Download Locally Grown Produce as a Marketing Strategy

University of Tennessee, Knoxville
Trace: Tennessee Research and Creative
Exchange
Masters Theses
Graduate School
5-2012
Locally Grown Produce as a Marketing Strategy:
Producer Perceptions of State-Sponsored
Marketing Programs
James Andrew Davis
[email protected]
Recommended Citation
Davis, James Andrew, "Locally Grown Produce as a Marketing Strategy: Producer Perceptions of State-Sponsored Marketing
Programs. " Master's Thesis, University of Tennessee, 2012.
http://trace.tennessee.edu/utk_gradthes/1147
This Thesis is brought to you for free and open access by the Graduate School at Trace: Tennessee Research and Creative Exchange. It has been
accepted for inclusion in Masters Theses by an authorized administrator of Trace: Tennessee Research and Creative Exchange. For more information,
please contact [email protected].
To the Graduate Council:
I am submitting herewith a thesis written by James Andrew Davis entitled "Locally Grown Produce as a
Marketing Strategy: Producer Perceptions of State-Sponsored Marketing Programs." I have examined the
final electronic copy of this thesis for form and content and recommend that it be accepted in partial
fulfillment of the requirements for the degree of Master of Science, with a major in Agricultural
Economics.
Margarita Velandia, Major Professor
We have read this thesis and recommend its acceptance:
Dayton M. Lambert, Christopher D. Clark, Michael D. Wilcox Jr., Kimberly Jensen
Accepted for the Council:
Dixie L. Thompson
Vice Provost and Dean of the Graduate School
(Original signatures are on file with official student records.)
Locally Grown Produce as a Marketing Strategy: Producer Perceptions of State-Sponsored
Marketing Programs
A Thesis
Presented for the
Master of Science Degree
The University of Tennessee, Knoxville
James Andrew Davis
May 2012
Abstract
State programs promoting their own agricultural products have proliferated in response to
increased consumer interest in locally grown foods (LGF). Tennessee, for example, currently has
two state- funded programs promoting the consumption of Tennessee agricultural products: Pick
Tennessee Products (PTP) and Tennessee Farm Fresh (TFF). The goal of this study is to examine
the factors affecting producer awareness and participation in these state-sponsored marketing
programs. This goal was achieved using survey data gathered from Tennessee’s fruit and
vegetable producers. These results should interest individuals attempting to increase producer
awareness and participation in these types of programs.
This thesis examines both producer awareness and participation in state-sponsored
marketing programs. The first essay of the thesis focuses on factors affecting Tennessee fruit and
vegetable producer awareness of TFF and PTP. The second essay examines factors that affect
Tennessee fruit and vegetable producer participation in TFF and PTP.
The factors affecting producer awareness of Tennessee’s two state-sponsored marketing
programs were evaluated using a bivariate probit model. Factors used in the analysis included
observed producer, farm, and regions characteristics. Findings suggest that producer awareness
was associated with education, percentage of income from farming, use of University/Extension
publications, attendance at University/Extension education events, and operation location.
A bivariate probit model was used to examine the effect of observed producer, farm, and
county characteristics on producer participation in TFF and PTP, given awareness of these
ii
programs. Results suggest that farmer participation in these programs was associated with size of
operation, education, use of Extension resources, and sale of fresh produce.
iii
TABLE OF CONTENTS
Part 1: Introduction ......................................................................................................................... 1
Introduction ................................................................................................................................. 2
Literature Review ........................................................................................................................ 3
Producers Perceptions of State Sponsored Marketing Program .............................................. 3
Consumers Perceptions of State Sponsored Marketing Program ............................................ 4
Objectives ................................................................................................................................ 9
Thesis Outline .......................................................................................................................... 9
References ................................................................................................................................. 10
Part 2: Factors Affecting Producer Awareness of State Programs Promoting Locally Grown
Foods: The Case of Fruit and Vegetable Growers in Tennessee .................................................. 12
Abstract ..................................................................................................................................... 13
Introduction ............................................................................................................................... 13
Description of Data ................................................................................................................... 16
Empirical Model........................................................................................................................ 17
Estimation Methods................................................................................................................... 21
Multicollinearity Tests ........................................................................................................... 21
Results and Discussion.............................................................................................................. 23
Sample Overview and Descriptive Statistics ......................................................................... 23
Bivariate Probit Marginal Effects .......................................................................................... 27
Conclusions ............................................................................................................................... 29
References ................................................................................................................................. 31
Appendix ................................................................................................................................... 34
Part 3: Factors Affecting Producer Participation in State Programs Promoting Locally Grown
Foods: The Case of Fruit and Vegetable Growers in Tennessee .................................................. 42
Abstract ..................................................................................................................................... 43
Introduction ............................................................................................................................... 43
Data and Methodology .............................................................................................................. 45
Theoretical Analysis .............................................................................................................. 46
Empirical Model .................................................................................................................... 50
Results and Discussion.............................................................................................................. 55
Sample Overview and Descriptive Statistics ......................................................................... 55
Evaluation of Factors Affecting Conditional and Joint Probabilities of Participation in TFF
and PTP: A Bivariate Probit Model ....................................................................................... 57
Conclusions ............................................................................................................................... 61
References ................................................................................................................................. 63
Appendix ................................................................................................................................... 65
Part 4: Summary............................................................................................................................ 76
Summary ................................................................................................................................... 77
Vita ............................................................................................................................................ 81
iv
LIST OF TABLES
Part 2: Factors Affecting Producer Awareness of State Programs Promoting Locally Grown
Foods: The Case of Fruit and Vegetable Growers in Tennessee
Table 1. Definitions and Descriptive Statistics of Variables (n=316) .......................................... 34
Table 2. Variable means for respondents aware of the Tennessee Farm Fresh and Pick Tennessee
Products programs and those not aware of the programs. ............................................................ 36
Table 3. Parameter Estimates from the Bivariate Probit Models for Estimating Factors Affecting
Awareness of Tennessee Farm Fresh and Pick Tennessee Products ............................................ 37
Table 4. Conditional Marginal Effects from the Bivariate Probit Model for Estimating Factors
Affecting Awareness of Tennessee Farm Fresh and Pick Tennessee Products. ........................... 39
Part 3: Factors Affecting Producer Participation in State Programs Promoting Locally Grown
Foods: The Case of Fruit and Vegetable Growers in Tennessee
Table 5. Definitions and Descriptive Statistics of Variables Affecting Awareness of Tennessee
Farm Fresh (n=301) ...................................................................................................................... 65
Table 6. Definitions and Descriptive Statistics of Variables Affecting Awareness of Pick
Tennessee Products (n=302) ......................................................................................................... 67
Table 7. Definitions and Descriptive Statistics of Variables Affecting Participation in TFF and
PTP (n=104) .................................................................................................................................. 69
Table 8. Variable Means for Respondents Participating in the Tennessee Farm Fresh, Pick
Tennessee Products, and Both Programs and Those Not Participating. ....................................... 70
Table 9. Heckman Sample Selection Model Estimation of Participation in the Tennessee Farm
Fresh and Pick Tennessee Products Program Given Awareness of These Programs. .................. 71
Table 10. Parameter Estimates from the Bivariate Probit Models for Estimating Factors
Affecting Participation in Tennessee Farm Fresh and Pick Tennessee Products ......................... 73
v
LIST OF FIGURES
Part 2: Factors Affecting Producer Awareness of State Programs Promoting Locally Grown
Foods: The Case of Fruit and Vegetable Growers in Tennessee
Figure 1. Comparison of Age Between Awareness Sample Respondents (n=316) and 2007
Census of Agriculture Data............................................................Error! Bookmark not defined.
Part 3: Factors Affecting Producer Participation in State Programs Promoting Locally Grown
Foods: The Case of Fruit and Vegetable Growers in Tennessee
Figure 2. Comparison of Age Between Participation Sample Respondents (n=316) and 2007
Census of Agriculture Data............................................................Error! Bookmark not defined.
vi
Part 1: Introduction
1
Introduction
Sales of locally grown food (LGF) in the U.S. grossed to $4.8 billion in 2008 and is
expected to grow to $7 billion by 2012 (Low and Vogel 2011; USDA/Agricultural Marketing
Services). U.S. consumer preferences for these products are driven by demand for freshness,
support of local economies, information about the source of the products, and reduction of the
environmental impact of the food chain when buying LGF (Food Marketing Institute 2009). In
particular, consumers perceive purchasing LGF as a way to reduce the environmental impact of
foods being transported long distances (Food Marketing Institute 2009). Consumers may also
associate sustainable methods of agricultural production with LGF and relate LGF with
production practices that reduce or eliminate the use of chemicals, moderate the impact of
agriculture on soil quality, water and air pollution (Thompson et al. 2008). Additionally,
consumer interest in knowing who produces the food may be a contributing factor to the increase
in popularity of LGF. The “story behind the food,” or “provenance” is often comprised of factors
such as who produced the food, the personality and ethics of the producer, and the attractiveness
of the farm or surrounding area (Thompson et al. 2008).
Although consumer interest in LGF has grown in recent years, s tates’ effort to promote
agricultural products grown within their limits is not new. States have been promoting their
products since about the 1930’s (Patterson 2006). Nonetheless, due to recent increase in LGF
popularity, state marketing programs have steadily increased (Onken and Bernard 2010). About
56% of all state marketing programs in the U.S. were established in 2000 or later (Onken and
Bernard 2010). In Tennessee, there are currently two state-funded programs to support and
develop markets for Tennessee-grown products. Pick Tennessee Products (PTP) was created by
2
the Tennessee Department of Agriculture (TDA) in 1986. In 2008, the Tennessee Department of
Agriculture, in cooperation with the Tennessee Farm Bureau, also created Tennessee Farm Fresh
(TFF). The purpose of these programs is to help farmers market their local products and inform
consumers about LGF markets.
Previous studies have looked at consumer awareness and preferences for products labeled
with state program logos and the impact of state marketing program on premiums, sales, and
local economies in general (Brooker and Eastwood 1989; Adelaja, Brumfield, and Lininger
1990; Govindasamy et al. 1998b; Patterson et al. 1999; Govindasamy et al. 2004; Carpio and
Isengildina-Massa 2010; Onken and Bernard 2010; Onken, Bernard, and Pesek 2011). Although,
these studies have evaluated the effectiveness of state marketing programs by measuring
consumer awareness (e.g. Govindasamy et al. 1998b; Patterson et al. 1999; Onken and Bernard
2010; Onken, Bernard, and Pesek 2011) or evaluating the impact of state marketing programs on
producer revenues or surplus (Brooker and Eastwood 1989; Govindasamy et al. 2004; Carpio
and Isengildina-Massa 2010), none of them have evaluated producer awareness and participation
in these programs. Little research exists of producer response to state-sponsored promotion
programs (Govindasamy et al. 1998a).
Lite rature Review
Producers Perceptions of State Sponsored Marketing Program
Govindasamy et al. (1998a) evaluated farmer awareness, participation, perceptions and
opinions about the Jersey Fresh logos (i.e., Jersey Fresh, quality grading, and premium logos)
using a survey of New Jersey farmers. About 93% of producers were aware of the Jersey Fresh
Program, and about 51% have used the Jersey Fresh logos. The majority of farmers (i.e., 91%)
3
indicated they had used Jersey Fresh logos to add locally grown value to their produce. About 37
of respondents perceived their average gross sales have grown, while about 43% indicated they
did not know whether the Jersey Fresh logos had any impact on sales. This study used logistic
regressions to identify factors affecting farmer use, and willingness to use Jersey Fresh logos in
the future. Results suggest that producers who perceived high consumer awareness of Jersey
Fresh logos, who used other logos to identify fresh produce, with more farming experience, and
located in the agricultural zone of New Jersey were more likely to use Jersey Fresh logos. In
contrast, farmers with larger acreage and more than 75% of their production being wholesaled
were less likely to use these logos. Additionally, results suggested that younger and more
educated farmers were more likely to be willing to participate in the Jersey Fresh Program in the
future. The profile of farmers who had the potential to use Jersey Fresh logos in the future was
very similar to the one of farmers already using the logos.
Consumers Perceptions of State Sponsored Marketing Program
Brooker and Eastwood (1989) use a survey of consumers in Knox County, Tennessee to
explore the impact of state sponsored logos on LGF sales in retail stores. They found that
consumers in Knox County had positive attitudes towards state logos. Consumer perceives state
logos as useful tools to identify LGF in metropolitan supermarkets, more for fresh products than
for processed food. Nonetheless, a small proportion of consumers were willing to pay a premium
for those products labeled with state sponsored logos.
Adelaja, Brumfield, and Lininger (1990) evaluated the potential success of the “Jersey
Fresh” brand.” A consumer survey was conducted at four stores in Northern New Jersey in 1988
to assess produce purchaser response to “Jersey Fresh” brand, specifically “Jersey Fresh”
4
tomatoes. Econometric methods (i.e., GLS and Iterative Zellner’s seemingly unrelated regression
techniques) were used to estimate demand functions for tomatoes grown in New Jersey using
survey cross-sectional data. They found that the Jersey fresh tomato seems to have a more
inelastic demand with respect to price, a more elastic demand with respect to income, and fewer
substitutes when compared to other products. Additionally, they concluded that consumer
preferences for Jersey Fresh tomatoes were based on the tomatoes’ quality and local origin. This
study found that characteristics of the demand function as well as consumer preferences for high
quality tomatoes grown in New Jersey represented a great opportunity for an increase in market
share of products grown in New Jersey, and therefore a potential success of the “Jersey Fresh”
brand promotion campaign.
Govindasamy, Italia, and Thatch (1998b) evaluated effectiveness of the Jersey Fresh
program through consumer awareness. A survey was conducted between July and August of
1996 to collect information about New Jersey consumer opinions about locally grown produce,
relative importance they place on price, quality and freshness when purchasing products, and
awareness and opinions about the Jersey Fresh campaign. They used a logistic regression to
identify factors affecting consumer awareness of the Jersey Fresh program. This study found
high awareness of the Jersey Fresh program among consumers. They found that consumers who
were more likely to be aware of this program shopped at direct marketing outlets, read food
advertisements, shop at more than one place, have lived in New Jersey for more than five years,
and own a farm garden.
Patterson et al. (1999) evaluated effectiveness of the Arizona Grown campaign. Using a
survey of grocery shoppers they evaluated consumer awareness and preferences for Arizona
5
Grown products. A two stage probit estimation procedure was used to identify factors affecting
consumer awareness and preferences for this program. They found limited awareness of the
Arizona Grown program among grocery shoppers and that the campaign had a fairly small
impact on consumer preferences for products grown in Arizona. Additionally, this study
collected information of product sales from a sample of retail grocers to evaluate impact of the
Arizona Grown program on sales. Using this information, this study estimated demand equations
for ten fruits and vegetables products promoted by the state marketing program. Results
suggested modest or no impact of the Arizona Grown campaign on in-store sales.
Govindasamy et al. (2004) evaluated the impact of the Jersey Fresh program on cash
receipts of farmers and the local economy in general. They estimated a promotional response
function to determine the impact of expenditures in the Jersey Fresh campaign on fruit and
vegetables producer cash receipts using an ordinary least square model. They concluded that
every dollar invested in the campaign increased New Jersey fruit and vegetable farmer revenues
by $31.54 and related industry revenues by $22.95. Therefore it was estimated that the $1.16
million investment in the Jersey Fresh campaign in 2000 generated $36.6 million in revenues for
New Jersey fruit and vegetable growers and $26.6 million in revenues for related industries for a
total impact of $63.2 million.
Onken and Bernard (2010) examined effectiveness of state branding programs in five
Mid-Atlantic States (New Jersey, Virginia, Maryland, Delaware, and Pennsylvania) though the
evaluation of consumer awareness of these programs. Using information from a consumer survey
conducted in the fall of 2009, they found that states with programs established in the 1980’s,
New Jersey and Virginia, had significantly higher awareness rates among survey respondents.
6
They also observed that, with the exception of New Jersey, consumers purchased more food
products labeled as “locally grown” as opposed to products labeled with the state program. They
suggested two potential causes for this result: 1) consumers cannot find or are unaware of
products advertised through state’s program; 2) consumers are more concerned with the concept
of “local” and they define this concept differently than the borders of the state. The lack of
consumer awareness of and the inability to find products label with the state program may
provide sales opportunities for state marketing programs (Onken and Bernard 2010).
Carpio and Isengildina-Massa (2010) provided a novel approach for an ex ante evaluation
of regional promotion campaigns. They used an equilibrium displacement model to evaluate the
impact of regional promotion campaign on quantities and prices. Additionally, they evaluated the
shift in demand for branded products due to a regional campaign using a contingent valuation
technique. This approach was applied to South Carolina’s locally grown campaign to measure
potential impact and effectiveness of this program. Using consumer surveys conducted in 2007
they evaluated consumer willingness to pay for produce and animal products grown in South
Carolina versus products grown out of the state. They concluded that the South Carolina locally
grown campaign increased consumers’ willingness to pay for produce by 3.4%, it increased
producer surplus by about three million dollars, and it represented a return to investment of about
618% for the state of South Carolina.
Onken, Bernard, and Pesek (2011) evaluated consumer preference and willingness to pay
for attributes such as organic, natural, locally grown, and products labeled under state marketing
programs. Using a choice experiment of five Mid-Atlantic States - New Jersey, Virginia,
Maryland, Delaware, and Pennsylvania - they determine willingness to pay for attributes in
7
strawberry preserves as well as the influence of purchasing venue on willingness to pay. They
found that consumers preferred local and state program preserves over non- local. Consumers
from New Jersey were the only ones who expressed preferences for state program over local
preserves. Consumers from larger states (i.e., Maryland, Pennsylvania, and Virginia) exhibit
preferences of local preserves over state program preserves. They found higher price premiums
for products labeled as local or state program promoted at farmers markets compared to grocery
stores settings. They concluded that state marketing program will have to be evaluated for
effectiveness by looking at consumers preferences for “local” compared to state program
promoted labels. This information may help identify whether or not a program is worth
continuing based on the potential premiums associated with state programs promoted brands.
Given the limited literature concerning marketing program awareness and participation
from a producer stand point this study intends to add to the existing literature about this topic. A
first step in evaluating effectiveness of these programs from producers’ perspectives is to better
understand awareness of the programs among those producers who would be most likely to
benefit from the services offered by state marketing programs. A second step in gauging
effectiveness of the programs is to look at actual participation in these programs. Information
about factors affecting producer awareness and participation in state marketing programs may be
of assistance to policy makers such as the Tennessee Department of Agriculture and
organizations that operate similar programs in other states, as well as University/Extension
personnel to expand the marketing potential of these programs. Additionally, this approach may
give a different perspective about how producers have benefited from state marketing
campaigns, and who else may have benefited from these programs. Finally, this information may
8
also help policy makers adjust limited funds to better promote state marketing programs and
increase participation across the state given that financial support for these programs from state
and private sources has been relatively modest and variable over time (Patterson 2006).
Objectives
The objectives of this research are: 1) to assess awareness of state marketing programs
(i.e., TFF and PTP) among Tennessee’s fruit and vegetable producers and to identify and
evaluate the factors associated with producer awareness, and 2) to evaluate factors affecting
Tennessee’s produce farmer participation in TFF and PTP programs.
Thesis Outline
The objectives of this thesis will be addressed in two essays. In the first essay, the extent
of awareness of TFF and PTP among Tennessee’s fruit and vegetable producers will be
evaluated. The factors affecting awareness of the two programs will then be examined using a
bivariate probit model. Producer participation in TFF and PTP and factors affecting participation
will be evaluated in the second essay using a similar approach.
The thesis will be organized as follows: part two presents description of data, empirical
model, estimation methods, results and discussion, and conclusions for the first essay. Part three
presents data and methodology, results and discussion, and conclusions for the second essay.
Note that in parts two and three a general introduction to the problem and specific objectives are
presented. Finally, in part four the two essays will be summarized and concluding comments
provided.
9
References
Adelaja, A.O., R.G. Brumfield, and K. Lininger. 1990. Producer Differentiation and State
Promotion of Farm Produce: An Analysis of the Jersey Fresh Tomato. Journal of Food
Distribution Research 21(3): 73-85.
Brooker, J.R., and D.B. Eastwood. 1989. Using State Logos to Increase Purchase of Selected
Food Products. Journal of Food Distribution Research 20(1): 175-183.
Carpio, C.E. and O. Isengildina-Massa. 2010. To Fund or Not to Fund: Assessment of the
Potential Impact of a Regional Promotion Campaign. Journal of Agricultural and
Resource Economics 35(2): 245-260.
Food Marketing Institute. 2009. U.S. Grocery Shopper Trends. Food Marketing Institute:
Arlington, VA.
Govindasamy, R., A. Pingali, J. Italia, and D. Thatch. 1998a. Producer Response to StateSponsored Marketing Programs: The Case of Jersey Fresh. Department of Agricultural,
Food and Resource Economics Rutgers University, P Series 02137-3-98.
Govindasamy, R., B. Schilling, K. Sullivan, C. Turvey, L. Brown, and V. Puduri. 2004. Returns
to the Jersey Fresh Promotional Program: The Impacts o f the Promotional Expenditures
on Farm Cash Receipts in New Jersey. Working paper, Department of Agricultural, Food
and Resource Economics, and the Food Policy Institute, Rutgers, The State University of
New Jersey.
Govindasamy, R., J. Italia and D .Thatch. 1998b. "Consumer Awareness of State-Sponsored
Marketing Programs: An Evaluation of the Jersey Fresh Program." Journal of Food
Distribution Research. 29(1998 d): 7-15.
Onken, K.A., and J.C. Bernard 2010. Catching the “Local” Bug: A Look at State Agricultural
Marketing Programs. Choices 25(1). Internet site:
http://www.choicesmagazine.org/magazine/article.php?article=112 (Accessed March 8,
2012).
Onken, K.A., J.C. Bernard, and J.D. Pesek Jr. 2011. Comparing Willingness to Pay for Organic,
Natural, Locally Grown, and State Marketing Program Promoted Foods in the MidAtlantic Region. Agricultural and Resource Economics Review 40 (1): 33-47.
Patterson, P.M. 2006. State-Grown Promotion Programs: Fresher, Better? Choices 21(2006): 4146.
Patterson, P.M., H. Oloffson, T. J. Richards, and S. Sass. 1999. An Empirical Analysis of State
Agricultural Product Promotions: A Case Study of Arizona Grown. Agribusiness: An
International Journal 15: 179-196.
10
Thompson, E., Jr., A.M. Harper, and S. Kraus. 2008. Think Globa lly- Eat Locally: San Francisco
Foodshed Assessment, American Farmland Trust. Internet site: http://www.farmland.org/
programs /states/ca/Feature%20Stories/San-Francisco-Foodshed-Report.asp (Accessed
February 1, 2012).
11
Part 2: Factors Affecting Producer Awareness of State Programs Promoting
Locally Grown Foods: The Case of Fruit and Vegetable Growers in Tennessee
12
Abstract
Interest in locally grown foods has increased over the past few years. Tennessee currently
has two state-funded programs promoting the consumption of Tennessee agricultural products by
linking producers and consumers-Tennessee Farm Fresh and Pick Tennessee Products. Factors
associated with fruit and vegetable producer awareness of each of these programs are analyzed
using a bivariate probit model. Findings suggest that awareness was associated with education,
percentage of income from farming, use of University/Extension publications, attendance at
University/Extension education events, and operation location. These results should be of
assistance to individuals attempting to increase producer awareness of programs promoting
locally grown foods.
Introduction
Interest in locally grown foods (LGF) has dramatically increased over the past few years.
In 2008, the U.S. market for LGF reached $5 billion (Tropp 2008). Big box retailers and grocery
chains increasingly dedicate shelf space to differentiate “locally grown” from “conventional”
produce as evidenced by Wal- mart, the top buyer of LGF at $400 million (Gambrell 2008).
Interest in Community Supported Agriculture (CSA) programs is also growing (Brown and
Miller 2008), and farmers markets are flourishing. Between 2000 and 2006, the number of
farmers markets increased by 8.6% per year to 4,093 nationwide (Agricultural Market ing Service
USDA). In Tennessee, the number of farmers participating in direct farm sales to consumers
increased by 33% from 1997 to 2007. The number of farmers markets in Tennessee increased by
56% from 2006 to 2009.
13
There are several reasons for the increased interest in LGF (Onken, Bernard, and Pesek
2011). LGF may provide health and nutrition benefits because they may be fresher and their
increased availability may encourage consumers to make healthier food choices (Martinez et al.
2010). LGF may also play a role in ameliorating a community’s concerns over food security 1 .
LGF provide a way for consumers to support local farmers and local economies (Gregoire and
Strohbehn 2002; Peterson, Selfa, and Janke 2010; Starr et al. 2003). The sales retained within a
region as consumers substitute LGF for imported products increases local farm revenue and
regional income (Swenson 2009). Finally, consumption of LGF may have environmental
benefits in reducing food miles to market, thereby moderating the use of fossil fuels in
transportation (Anderson 2007; Gomez 2010) 2 .
Because of these perceived benefits, federal and state governments have adopted a
number of programs to support producers attempting to supply LGF (Martinez et al. 2010;
Onken, Bernard, and Pesek 2011). Examples of federal programs include the Fresh Program, the
Women and Infant Childcare (WIC) Farmers Market Nutrition Program (FMNP), and the
Senior’s Farmers Market Nutrition Program (SFMNP). The Fresh Program is a partnership of the
U.S. Department of Defense and the U.S. Department of Agriculture (USDA) to promote the
consumption of fresh, locally grown foods by schools and other institutions. The FMNP and
SFMNP issue coupons to seniors and WIC participants that can be used at authorized farmers
markets, roadside stands, and CSAs.
1
Food security has been defined as all people at all t imes having access to enough food for an active, healthy life
(Nord and Andrews, 2002).
2
The extent to which a shift toward LGF would actually engender environ mental benefits is uncertain given that
distance traveled is an imperfect measure of the environ ment impact of food transportation (Co ley, Ho ward, and
Winter 2009) and that the production of food typically has a larger impact on the environ ment than its transportation
(Weber and Matthews 2008).
14
There are also a number of state- level programs designed to promote the consumption of
LGF. For example, in Tennessee there are currently two state- funded programs to support and
develop markets for Tennessee-grown products. Pick Tennessee Products (PTP) was created by
the Tennessee Department of Agriculture in 1986. In 2008, the Tennessee Department of
Agriculture - this time in cooperation with the Tennessee Farm Bureau - created Tennessee Farm
Fresh (TFF). The purpose of both programs is to link producers with marketing channels for
LGF and to inform consumers about opportunities to purchase LGF. The PTP program promotes
all products available at Tennessee farms, farmers markets, and other retail outlets, while TFF
focuses on the promotion of fresh products grown in Tennessee, including fruit and vegetables,
nursery, dairy and some livestock products. The two programs offer an array of similar benefits,
including: a listing on a web-site directory, the right to use the TFF and PTP logos, and
advertising benefits. The two programs are differentiated by the following: the TFF program
offers a banner with the TFF logo, TFF stickers, price cards, TFF reusable bags, and free access
to workshops offered through the University of Tennessee Center for Profitable Agriculture to
their members while the PTP program offers the right to participate in their on- line store but
participation in this programs does not guarantee access to marketing tools (e.g. banner, price
cards, stickers, workshops) (Howard 2012). Additionally, there are no fees required to participate
in the PTP program, but the TFF program charges a $100 annual fee for participation.
A first step in gauging the effectiveness of these programs is to better understand
awareness of the programs among those producers who would be most likely to benefit from the
services offered by the two programs. Thus, the objectives of this study are to gauge awareness
of the programs among Tennessee’s fruit and vegetable producers and to identify and evaluate
15
the factors associated with producer awareness. The study’s focus is on fruit and vegetable
producers because produce growers account for a large portion of direct agricultural sales
(USDA 2007; Onken, Bernard, and Pesek 2011), which is one of the main marketing outlets for
LGF (Martinez et al. 2010; Low and Vogel 2011). The information provided by this study should
be of assistance to governmental agencies and other institutions that are interested in increasing
producer awareness of programs or other efforts promoting LGF. Greater awareness of such
programs or efforts may help producers increase profit margins through the adoption of new
marketing strategies.
Description of Data
This study uses data from a 2011 survey of Tennessee’s fruit and vegetable producers.
The list frame for the survey was provided by USDA’s National Agriculture Statistics Service
(NASS) and included the entire population of fruit and vegetable producers in Tennessee. On
February 2, 2011, the survey, a cover letter explaining the importance of the survey, and a
postage paid return envelope were sent to Tennessee’s 1,954 fruit and vegetable producers by
first class mail. Approximately three weeks later, reminder postcards were sent. One month later,
a second wave of surveys was mailed to those who had not returned the survey. Of the 1,954
questionnaires mailed, 587 were completed and returned, providing a response rate of
approximately 30%. After eliminating observations with missing data, 316 responses were
suitable for this analysis.
The survey included questions about: marketing outlets used to sell fruits and vegetables;
barriers producers faced when participating in different markets; perceptions of the
characteristics that define a “local” market; awareness of, and participation in, Tennessee’s
16
programs promoting LGF (i.e., TFF and PTP); and general farm business and operator
characteristics. Secondary data concerning food marketing and other environmental factors or
community characteristics (e.g. metro/non-metro county, number of farmers markets in a county)
were collected from the Food Environmental Atlas (http://www.ers.usda.gov/foodatlas, USDA,
2011).
Empirical Model
Produce grower awareness of the TFF and PTP programs can be empirically specified as,
y i1   1' xi1  ei1
(1)
y i 2   1' xi 2  ei 2
Corr( ei1 , ei 2 )  
where yi1 =1 if a producer is aware of TFF, and zero otherwise; yi2 =1 if a producer is aware of
PTP, and zero otherwise; β1 , and β2 are parameters associated with each awareness equation; ei1 ,
and ei2 , are random disturbances for each equation; and xi1 , and xi2 are vectors of observed
producer, farm, and county characteristics that may influence the likelihood that a producer is
aware of either program. Given the similarities in the two programs, there are unobserved
variables that are likely to similarly influence awareness of each of the programs and, thus, the
error terms for the two equations are likely to be correlated ( Corr( ei1 , ei 2 )   ). A description of
the variables used in this analysis is presented in Table 1.
Producer characteristics hypothesized to influence awareness of PTP and TFF are: age
(AGE); highest level of educational attainment, expressed in dicho tomous variables for some
high school (SOMEHS), high school graduate (HSGRAD), some college (SOMECOLL),
associate’s degree (ASSOCDEG), bachelor’s degree (BACHDEG), and graduate degree
17
(GRADDEG); the percentage of taxable household income coming from farming, expressed in a
dichotomous variable for less than 25 percent (PF_INCOME); the number of
University/Extension educational events or presentations related to produce marketing that the
grower had attended in the past five years (EDUC_EVENTS); and whether the producer had
used University/Extension publications to obtain information about how to better market produce
in the last 5 years (PUBLICATIONS).
Age is expected to be negatively correlated with awareness as older producers tend to
have shorter planning horizons and may be less likely to search for programs that offer
alternatives to current marketing efforts. Education is expected to be positively correlated with
awareness as marketing produce directly to consumers requires special skills and abilities, not all
of which are likely to be directly related to agricultural operations (Uva 2002; Uematsu and
Mishra 2011). Thus, given that direct marketing to consumers is one of the main marketing
outlets for LGF (Martinez et al. 2010; Low and Vogel 2011), it is expected that more educated
farmers may be more willing to experiment with LGF marketing strategies and more likely to be
aware of programs promoting LGF. The percentage of household income from farming is
hypothesized to be positively correlated with awareness of the programs, as producers with a
high percentage of income from farming are more likely to be willing to invest the time and
effort needed to improve their bottom line sales through novel marketing strategies and,
therefore, more likely to be aware of programs designed to meet those needs. Attendance at
University/Extension outreach events or presentations related to produce marketing strategies is
expected to increase producer exposure to, and thus awareness of, the programs. Similarly, the
18
use of University/Extension publications to obtain information about how to better market
produce is also expected to increase producer awareness of these programs.
The characteristics of the producer’s operation included in the analysis are: size of the
producer’s fruit and vegetable operation in acres (VEGSIZE); percentage of sales made directly
to consumers (TDS), intermediaries (TIN), and retail outlets (TRE); percentage of direct sales to
consumers in different geographic areas, expressed in dichotomous variables for in: the
producer’s county of operation (YOURCNTY); neighboring counties (NEXTCNTY); elsewhere
in the state of Tennessee (INSTATE); elsewhere in the U.S. (INUS); and elsewhere in the world
(OTHCNTRY).
It is hypothesized that the size of the producer’s fruit and vegetable operation will be
negatively correlated with awareness of the two programs. Producers managing larger operations
may be more inclined to market products through wholesalers, whereas smaller operations might
rely more on alternative marketing channels such as farmers markets and CSAs (Lockeretz 1986;
Low and Vogel 2011; Watson and Gunderson 2010) where the services provided by the two
programs would be of more use.
The percentage of sales made directly to consumers is likely to be positively correlated
with producer awareness of the PTP and TFF programs as the services offered by these programs
would seem to be more directly applicable to these types of sales. In addition, it could be that the
concept of “local” is more important to the consumers who purchase produce directly from
producers (Lockeretz 1986). Similarly, farmers who market produce directly to consumers
through farmers markets and CSAs may have a greater chance of being exposed to programs
19
promoting LGF as other producers also selling through these outlets may be already participating
in programs promoting LGF (Low and Vogel 2011). Producers who market a greater share of
their produce through intermediaries (e.g. wholesalers, grower cooperatives) or retailers (e.g.
groceries) are less likely to be aware of programs promoting LGF, because the services offered
by these programs may be less relevant to these types of sales and because consumers who
purchase their produce through these outlets might be more interested in price than other
characteristics (Lockeretz 1986). The percentage of a producer’s direct sales to consumers in
Tennessee is likely to be positively correlated with awareness of the programs promoting LGF
given that the goal of these programs is to promote Tennessee-grown products. Therefore it is
hypothesized that producers with a larger percentage of sales elsewhere in the U.S. and other
countries are less likely to be aware of these programs.
The characteristics of the county in which the producer operates that are included in this
analysis are: whether the county is located in east (EASTTENN), middle (MIDTENN), or west
(WESTTENN) Tennessee; whether the county is a metropolitan county (METRO); and the
number of farmers markets operating in the county (FMRKT10). Geographic location could
influence producer awareness in a number of ways. Direct-to-consumer sales drivers are affected
by regional characteristics such as proximity to farmers markets and to farmland (Low and Vogel
2011). Therefore, geographic location may explain producer exposure to programs promoting
LGF. It is hypothesized that producers located in regions producing more fruit and vegetables
and other specialty crops, and closer to farmers markets and farmer-to-grocer’s marketing
channels are more likely to be aware of programs promoting LGF (Low and Vogel 2011). Thus,
it is also expected that the number of farmers markets located in the producer’s county will
20
positively influence the likelihood of program awareness. The greater the number of farmers
markets in a county the more likely farmers would be to market fresh produce to this outlet.
Given that farmers markets are one of the most popular direct to consumer outlets for LGF it is
expected that the greater the number of farmers markets in a county the more likely farmers are
to be exposed to programs promoting LGF (Low and Vogel 2011).
Estimation Methods
The error terms in the awareness equations presented in (1) are assumed to be normally
distributed and correlated ( Cov(e i1 , e i 2 )   ). Therefore a bivariate probit model based on the
join distribution of the error terms ( ei1 , ei 2 ) is used for this analysis. To construct the likelihood
function for this model let qi1  2 y i1  1 and qi 2  2 y i 2  1 . Thus
(2)
 1 if y im  1
q im  
m  1,2
 1 if y im  0
' β m , wim  q im z im , and  *  q i1 q i 2  .
Additionally, let z im  x im
i
The probabilities entering the likelihood function are (Green 2003):
(3)
Pr ob(Y1  y i1 , Y2  y i 2 | x i1 , x i 2 )   2 (wi1 , wi 2 ,  i* ) ,
where Φ2 denotes the bivariate normal cumulative distribution function. Therefore, the loglikelihood function can be defined as:
n
(4)
log L   ln  2 ( wi1 , wi 2 ,  i* )
I 1
The derivatives of the log- likelihood with respect to the parameters of interest (i.e., βim, ρ) are:
21
(5)
 ln L n  q im g im 
 
x im for m  1,2
β m

i 1   2
(6)
 ln L n q i1 q i 2 2

for m  1,2

2
i 1
where 2 denotes the bivariate normal density function and
(7)
 wi 2   i* wi1 
g i1   ( wi1 ) 
,
 1   i*2 
where  represents the univariate standard normal density and Φ represents the univariate
standard normal cumulative distribution function. The subscripts 1 and 2 are reverse in (7) to
obtain gi2 .The maximum likelihood estimates are obtained by simultaneously setting (5) and (6)
equal to zero. If   0 then  i*  0 and therefore
g i1   (wi1 )wi 2  .
(8)
Replacing (8) in (5) reduces to the first order condition of a probit model.
Marginal effects are computed given the bivariate nature of the model (Greene 2003). The
approach taken here was to first obtain the expected value of awareness of one of the programs
(say, yi1 =1), conditional on the respondent being aware of the alternative program (yi2 =1):
(9)
E ( y i1 y i 2  1, x)  Prob( y i1  1 | y i 2  1, x) 

 2 ( x' γ 1 , x' γ 2 ,  )
,
 ( x'  2 )
22
Prob( y i1  1, y i 2  1 | x)
Prob( y i 2  1 | x)
where x=x1  x2 , x’γm = x1 ’βm . Therefore γ1 contains all the nonzero elements of β1 and possible
some zeros in the positions of variables in x that appear only in equation 2 in (1). The derivative
of (9) was taken with respect to the explanatory variables of interest to estimate the marginal
effects:
E ( y i1 y i 2  1, x)
(10)
x
 1
 
  ( x'  2

 ( x'  2 )  
  g 1 1  ( g 2   2
 2  ,
 (x'  2 )  

where g1 and g2 are defined in (7).
Multicollinearity Tests
Multicollinearity may compromise inferences by inflating variance estimates (Greene
2003; Judge et al. 1988). A condition index was used to detect collinear relationships (Belsley,
Kuh, and Welsch 1980). Condition indexes between 30 and 100 indicate that the explanatory
variables have moderate to strong association with each other. A condition index accompanied
by a proportion of variation above 0.5 indicates potential collinearity problems (Belsley, Kuh,
and Welsch 1980).
Results and Discussion
Sample Overview and Descriptive Statistics
The average age of respondents included in this analysis (n=316) was 61 years, close to
the average farmer age in Tennessee (58 years) according to the 2007 Census of Agriculture
(USDA/NASS). The age distribution from respondents follows closely age distribution among
vegetable and melon farmers, and fruit and nut farmers in Tennessee (Figure 1). The proportions
of farmers in each age category are similar when comparing respondents age and age of
23
vegetable and melon farmers in Tennessee. The sample used in this study had a larger proportion
of farmers in the 35 to 44, 45 to 54 and 55 to 64 age categories when compared to Tennessee
fruit and nut farmers. However, the proportion of Tennessee fruit and nut farmers in the 65 and
over category was larger compared to the sample respondents. For about 26% of the respondents
the highest level of educational attainment was a bachelor’s degree, followed by 22% who
earned a graduate degree and 22% who graduated from high school but did not attend college.
About 69% of respondents earned less than 25% of their household income from farming.
Respondents had attended an average of 1.2 University/Extension educational events or
presentations related to marketing strategies over the past five years. About 30% of the
respondents had used University/Extension publications to obtain information about improving
their produce marketing within the past five years.
The average size of the fruit and vegetab le operations was 10.8 acres. The majority
(about 84%) of sales made by the respondents were direct sales to consumers. Most (about 69%
on average) of the direct sales made by the respondents in 2010 took place in their home county.
The average percentage of direct sales made in neighboring counties and elsewhere in the state
were 24% and 5%, respectively. About 42% of the respondents were located in Middle
Tennessee, 40% in East Tennessee, and the reminder in West Tennessee. About 47% of the
respondents lived in metropolitan counties.
About 42% of the respondents included in this analysis were aware of the TFF program
and 54% were aware of the PTP program. Greater awareness of the PTP program is probably not
too surprising given that it has been in existence for about 22 years longer than the TFF program.
Comparisons of the mean values for producer, producer operation and county characteristics, on
24
the basis of awareness of the TFF and PTP programs, are presented in Table 2. Differences in
mean values between those who were aware and those who were not aware of each program
were compared using t-tests. Significant differences for the variables associated with producer
characteristics were evaluated. The proportion of producers with 25% or less of their household
income from farming who were unaware of the TFF and PTP programs was larger (80% and
83%, respectively) than the proportion of producers with 25% or less of their income from
farming who were aware of these programs (55% and 58%, respectively). As expected,
producers with a higher percentage of income from farming are more likely to be aware of
programs design to increase sales through alternative marketing strategies, given that they have a
higher dependence on the economic viability of the farming operation. On average, respondents
who were aware of TFF and PTP had attended more University/Extension educational events or
presentations related to produce marketing over the last five years (2.1 and 1.9 events,
respectively) compared to respondents who were not aware of the programs (0.5 and 0.3,
respectively); as hypothesized, producers who attend these educational events may be more
interested in alternative produce marketing strategies and more likely to be exposed to
information about programs promoting LGF. Finally, about 48% of the respondents aware of
TFF and 42% of those aware of PTP have used University/Extension publications to obtain
information about how to better market their produce within the last five years, which is
significantly higher than the 17% and 15% of producers not aware of TFF and PTP, respectively
who used University/Extension publications for this purpose. University/Extension publications
related to produce marketing strategies may include information about programs promoting LGF
and therefore producers using these publications are more likely to be aware of TFF and PTP.
25
Significant differences for the variables associated with characteristics of the producer’s
operation were also considered. The average size of the fruit and vegetable operations was larger
for respondents aware of TFF and those aware of PTP (17.1 and 14.3 acres, respectively) than
those who were unaware of the programs (6.6 and 7.0 acres, respectively). Contrary to the
hypothesis that local food marketing is more likely to occur on smaller operations (Martinez et
al. 2010), for this sample, it seems that larger operations are more likely to be aware of programs
promoting LGF in Tennessee. The average percentage of fruit and vegetable sales made in the
county in which a producer’s operation was located was significantly higher for producers
unaware of TFF and PTP (75% and 78%, respectively) compared to producers who were aware
of the two programs (60% for both). However, the average percentage of sales made in
neighboring counties and elsewhere in the State was significantly higher for producers who were
aware of TFF and PTP (29% and 30%, respectively for sales in neighboring counties, and 8%
and 7%, respectively for sales elsewhere in the State) than for those who were unaware of the
programs (20% and 17%, respectively for sales in neighboring counties and 3% for sales
elsewhere in the State). As expected, producers with relatively more sales in Tennessee are more
likely to be aware of programs promoting LGF given that the goal of these programs is to
promote products grown in Tennessee. Nonetheless, respondents selling a higher percentage of
their produce within their county of operation were less likely to be aware of TFF and PTP.
Finally, significant differences associated with the characteristics of the county in which
the grower operates were identified. About 54% of the producers who were aware of TFF live in
metropolitan counties while only 42% of the producers not aware of the program live in
26
metropolitan counties. This result is explained by the fact that marketing of LGF is more likely
to take place in metropolitan counties (Martinez et al. 2010).
Bivariate Probit Marginal Effects
The results of the bivariate probit model can be seen in Table 4. The correlation
coefficient between the residuals (ρ) was positive and statistically significant at the 1% level,
supporting the hypothesis that the error terms in the TFF and PTP awareness equations were
correlated, and also suggesting that the bivariate probit approach appears appropriate. A
likelihood ratio test for the overall significance of the model indicated the model was significant
at the 1% level.
The marginal effects of the bivariate probit model used to examine the factors affecting
awareness of the TFF and PTP programs are presented in Table 3. Five of the explanatory
variables had statistically significant marginal effects on awareness of the TFF program, given
that the producer was aware of the PTP program. These five variables were whether the producer
had some high school education (SOMEHS), whether the producer had used
University/Extension publications to obtain information about marketing produce within the past
five years (PUBLICATIONS), the size of the producer’s fruit and vegetable operation in acres
(VEGSIZE), the percentage of the producer’s total sales made directly to consumers (TDS), and
whether the producer’s operation was located in a metropolitan county (METRO). Although
these marginal effects were statistically significant some of them were very small in magnitude
(i.e., VEGSIZE, TDS). The results suggest that producers located in a metropolitan county are
18% more likely to be aware of the TFF program, and producers who used University/Extension
publication are 20% more likely to be aware of TFF, given that they are already aware of the
27
PTP program. The marginal effect associated with the education variable (SOMEHS) has a
positive sign. This result suggests that producers with some high school education tended to be
more likely to be aware of TFF than producers with bachelor degrees. This result runs counter
the hypothesis that more educated farmers are more likely to be aware of programs promoting
LGF. A possible explanation for this result is that more educated farmers may be more likely to
be employed part time off the farm and therefore may have less time to look for alternative
marketing opportunities such as LGF. Statistically significant conditional marginal effects for the
PTP awareness equation were those associated with age (AGE), education (SOMEHS),
percentage of total household income from farming activities (PF_INCOME), and the number of
University/Extension educational events or presentations related to produce marketing strategies
attended within the past five years (EDUC_EVENTS). Again, some of the statistically
significant marginal effects were very small in magnitude (i.e., AGE). The results suggest that,
given awareness of the TFF program, producers with some high school education are 35% less
likely to be aware of PTP than producers with bachelor degrees, while producers with less than
25% of their income coming from farming are 7% less likely to be aware of the PTP program
and, finally, attending an additional educational event increases the likelihood of being aware of
PTP by 2.5%.
In summary, producers who are already aware of the PTP program and who have used
University/Extension publications to obtain information about how to better market produce in
the last 5 years, operate larger fruit and vegetable operations, derive a higher percentage of their
sales from direct-to-consumer outlets, and are located in metropolitan counties are more likely to
be aware of the TFF program. On the other hand, younger, more educated producers, with more
28
than 25% of their household income from farming, who have attended more
University/Extension educational events or presentations related to marketing strategies to sell
produce in the past five years are more likely to be aware of the PTP program, given awareness
of the TFF program.
Conclusions
The marketing of LGF continues to grow in popularity. The goal of this study is to
evaluate fruit and vegetable producer awareness of the two Tennessee programs designed to
enhance LGF marketing opportunities – TFF and PTP. A bivariate probit regression was used to
measure the association between the characteristics of the producer, the producer’s operation,
and the county in which the producer’s operation is located and producer awareness of these
programs.
The factors affecting awareness of TFF and PTP programs differed between the two
programs. Use of University/Extension publications, size of the fruit and vegetable operation,
percentage of sales from direct-to-consumers outlets, and location in a metropolitan county all
significantly affected awareness of the TFF program. On the other hand, attendance at
University/Extension education events, age, education, and percentage of income from farming
were factors significantly affecting producer awareness of the PTP program. Policymakers such
as the Tennessee Department of Agriculture and organizations that operate similar programs in
other states, as well as University/Extension personnel may benefit from this information to
better market these programs. This information may also help policy makers adjust limited funds
to better promote these programs by better targeting their clientele and increasing awareness of
the programs across the state.
29
Attendance at University/Extension educational events or presentations related to
produce marketing and the use of University/Extension publicatio ns to obtain information about
how to better market their produce were significant factors affecting awareness of both the PTP
and TFF programs. These results suggest that the partnership between policy makers and
Extension may increase effectiveness in spreading the word about state programs promoting
LGF. Therefore, it may be important for policymakers to continue working with Extension to
increase producer awareness of state programs promoting LGF. Nonetheless, producers who are
unaware of the TFF and PTP programs may not be attending University/Extension educational
events or presentations related to marketing strategies to sell produce and/or using
University/Extension publications. Therefore, reaching these producers will require alternative
strategies.
30
References
Anderson, M.D. 2007. The Case for Local and Regional Food Marketing. Farm and Food Policy
research brief. Northeast-Midwest Institute. Washington, DC. Internet site: http:// www.
foodsystems- integrity.com/yahoo_site_admin/ assets/docs/ The_Case_for_ Local__
Regional_Food_Marketing.28645058.pdf (Accessed on February 7, 2011).
Belsley, D.A., E. Kuh, and R.E. Welsch. 1980. Regression Diagnostics: Identifying Influential
Data and Sources of Collinearity. New York, NY: John Wiley & Sons.
Brown C., and Miller, S. 2008. “The Impacts of Local Markets: A Review of Research on
Farmers Markets and Consumer Supported Agriculture (CSA).” American Journal of
Agricultural Economics 90(5): 1296-1302.
Coley, D., M. Howard, and M. Winter. 2009. Local food, food miles and carbon emissions: A
comparison of farm shop and mass distribution approaches. Food Policy 34(2): 150155.
Gambrell, J., 2008. “Wal-Mart Now Top Buyer Of Locally Grown Produce.” Transport Topics
Publishing Group, August 4, 2008, 2 pp.
Gomez, M. 2010. “Moving Local Foods from Farm to Consumer: Lessons from NYS Apples. ”
Research & Policy Brief Series 38: Unpaginated
Greene, W.H. 2003. Econometric Analysis. 5th Edition. Upper Saddle, NJ: Prentice Hall.
Gregoire, M. and C. Strohbehn. 2002. “Benefits and Obstacles to Purchasing Food From Local
Growers and Producers.” The Journal of Child Nutrition & Management 26 (1):
Unpaginated. Internet site: http://docs.schoolnutrition.org /newsroom/ jcnm/
02spring/gregoire/ (Accessed on October 22, 2010).
Howard, T. 2012. Personal communication, Tennessee Farm Fresh program, Tennessee Farm
Bureau Federation, Knoxville, TN, January 25.
Judge, G.G., R.C. Hill, W.E. Griffiths, H. Lutkepohl, and T. Lee. 1988. Introduction to the
Theory and Practice of Econometrics. New York, NY: John Wiley & Sons.
Lockeretz, W. 1986. “Urban Consumers’ Attitudes towards Locally Grown Produce.” American
Journal of Alternative Agriculture 1(2): 83–88.
Low, S.A., and Vogel, S. 2011. “Direct and Intermediated Marketing of Local Foods in the
United States.” Washington DC: U.S. Department of Agriculture, ERS Economic
Research Rep. 128, November.
31
Martinez, M., M. Hand, M. Da Pra, S. Pollack, K. Ralston, T. Smith, S. Vogel, S. Clark, L. Lohr,
S. Low, and C. Newman. 2010. “Local Food Systems: Concepts, Issues, and Issues.”
Washington DC: U.S. Department of Agriculture, ERS Economic Research Rep. 97,
May.
Nord, M. and M. Andrews. 2002. “Reducing Food Insecurity in the United States: Assessing
Progress Toward a National Objective.” Washington DC: U.S. Department ofAgriculture,
ERS Food Assistance and Nutrition Research Rep. 26-2, May.
Onken, K.A., J.C. Bernard, and J.D. Pesek Jr. 2011. “Comparing Willingness to Pay for Organic,
Natural, Locally Grown, and State Marketing Program Promoted Foods in the MidAtlantic Region.” Agricultural and Resource Economics Review 40 (1): 33-47.
Peterson, H. H., T. Selfa, and R. Janke. 2010. “Barriers and Opportunities for Sustainable Food
Systems in Northeastern Kansas.” Sustainability 2: 232-251.
Starr, A., A. Card, C. Benepe, G. Auld, D. Lamm, K. Smith, and K. Wilken. 2003. “Sustaining
local agriculture: Barriers and opportunities to direct marketing between farms and
restaurants in Colorado.” Agriculture and Human Values 20: 301-321.
Swenson, D. 2009. “Estimating the Production and Market-Value Based Impacts of
Nutritional Goals in NE Iowa.” Leopold Center for Sustainable Agriculture. Ames, IA.
Available at: http://www.leopold.iastate.edu/research/marketing_files/seiowa.pdf
[Accessed on February 7, 2011].
Tropp, D., 2008. “The Growing Role of Local Food Markets,” American Journal of Agricultural
Economics 90(5): 1310-1311.
Uematsu, H., and A. K. Mishra 2011. "Use of Direct Marketing Strategies by Farmers and Their
Impact on Farm Business Income." Agricultural and Resource Economics Review
40(1): 1 – 19.
U.S. Department of Agriculture, Agricultural Marketing Service-Marketing Services Division.
Number of Operating Farmers Markets: 1994-2010. Internet site:
http://www.ams.usda. gov/AMSv1.0/ams.fetchTemplateData.do?template=
TemplateS&nav ID=Wholesaleand FarmersMarkets&leftNav=Wholesaleand
FarmersMarkets&page=WFMFarmersMarketGrowth&description=Farmers%20Marke
t%20Growth&acct=frmrdirmkt (Accessed November, 2010).
U.S. Department of Agriculture, National Agricultural Statistics Service. 2007. Census of
Agriculture. Internet site: http://www.agcensus.usda.gov/Publications/2007/ index.asp
(Accessed October 10 , 2011).
32
U.S. Department of Agriculture, Economic Research Services. Food Environemmnt Atlas.
Internet site: http://www.ers.usda.gov/foodatlas/ (Accessed July 15, 2011).
Uva, W.-F.L. 2002. “An Analysis of Vegetable Farms’ Direct Marketing Activities in New York
State.” Journal of Food Distribution Research 33(1): 186–189.
Watson, J.A., and M. Gunderson. 2010. Direct Marketing of Specialty Crops by Producers: A
Price-Comparison between Farmers’ Markets and Grocery Stores. Selected paper
presented at the Southern Agricultural Economics Association Annual Meeting, February
6-9, Orlando, FL. Internet site: http://ageconsearch.umn.edu/bitstream/56512/2/
Direct%20Marketing%20of%20Specialty%20Crops%20by%20Producers.pdf (Accessed
December 7, 2011).
Weber, C.L., and Matthews, H.S. 2008. “Food-Miles and the Relative Climate Impacts of Food
Choices in the United States.” Environmental Science & Technology 42(10): 3508-3513.
33
Appendix
Table 1. Definitions and Descriptive Statistics of Variables (n=316)
Variable
Description
A. Dependent
Variables
=1 if farmer is aware of Tennessee Farm Fresh,
AWARE_TFF
zero otherwise
=1 if farmer is aware of Pick Tennessee Products,
AWARE_PTP
zero otherwise
B. Independent
Variables
AGE
Age of producer in years
=1 if some high school is the highest level of
SOMEHS
education attained by the farmer, zero otherwise
=1 if high school diploma is the highest level of
HSGRAD
education attained by the farmer, zero otherwise
=1 if some college is the highest level of education
SOMECOLL
attained by the farmer, zero otherwise
=1 if an associate’s degree is the highest level of
ASSOCDEG
education attained by the farmer, zero otherwise
=1 if a bachelor’s degree is the highest level of
BACHDEG
education attained by the farmer, zero otherwise
=1 if a graduate degree is the highest level of
GRADDEG
education attained by the farmer, zero otherwise
=1 if less than 25% of farmer household income
PF_INCOME
comes from farming
The number of educational events the farmer has
EDUC_EVENTS
attended in the past 5 years
=1 if the farmer has used University/Extension
PUBLICATIONS
publications in the past 5 years
Percent of direct sales to consumers in the county
YOURCNTY
where the farmer operates
Percent of direct sales to consumers in neighboring
NEXTCNTY
counties of where the farmer operates
Percent of direct sales to consumers elsewhere in
INSTATE
the state
Percent of direct sales to consumers elsewhere in
INUS
the country
Percent of direct sales to consumers in other
OTHCNTRY
countries
VEGSIZE
Size of fruit and vegetable operation in acres
Percent of direct sales obtained from direct to
TDS
consumer outlets
34
Mean
0.4114
0.5380
60.7089
0.0633
0.2152
0.1519
0.0949
0.2595
0.2152
0.6962
1.1416
0.2975
68.5158
23.8070
5.3212
1.7547
0.6013
10.8920
84.4842
Table 1. Continued
Variable
TIN
TRE
EASTTENN
MIDTENN
WESTTENN
FMRKT10
METRO
Description
Percent of direct sales obtained from direct to
intermediary outlets
Percent of direct sales obtained from direct to retail
outlets
=1 if the farmer is located in East Tennessee, zero
otherwise
=1 if the farmer is located in Middle Tennessee,
zero otherwise
=1 if the farmer is located in West Tennessee, zero
otherwise
The number of farmers markets in the county where
the farmer operates
=1 if farmer is located in a metropolitan county,
zero otherwise
35
Mean
7.9114
7.6044
0.3956
0.4241
0.1804
1.0475
0.4684
Table 2. Variable Means for Respondents Aware of the Tennessee Farm Fresh and Pick
Tennessee Products Programs and Those Not Aware of the Programs
Pick Tennessee Products
Tennessee Farm Fresh
Not Aware
Aware
Not Aware
Aware
Independent Variablesa
(n=186)
(n=130)
(n=146)
(n=170)
AGE
61.6129
59.4154
63.5069***
58.3059
SOMEHS
HSGRAD
0.0591
0.1989
0.0692
0.2385
0.0822
0.2192
0.0471
0.2118
SOMECOLL
0.1613
0.1385
0.1644
0.1412
ASSOCDEG
BACHDEG
*
0.1183
0.2473
0.0615
0.2769
0.1164
0.2055**
0.0765
0.3059
GRADDEG
0.2151
0.2154
0.2123
0.2176
PF_INCOME
EDUC_EVENTS
0.7957***
0.4839***
0.5538
2.0827
0.8288***
0.2877***
0.5824
1.8750
PUBLICATIONS
0.1720***
0.4769
0.1507***
0.4235
YOURCNTY
74.5699
***
59.8539
78.4041
***
60.0235
NEXTCNTY
INSTATE
20.1613**
3.2957**
29.0231
8.2192
16.5411***
3.1370**
30.0471
7.1971
INUS
1.3817
2.2885
1.8151
1.7029
OTHCNTRY
VEGSIZE
0.5914
6.5880***
0.6154
17.0500
0.1027
6.9802**
1.0294
14.2515
TDS
84.3172
84.7231
86.9726
82.3471
TIN
TRE
7.7527
7.9301
8.1385
7.1385
5.6233*
7.4041
9.8765
7.7765
EASTTENN
0.4032
0.3846
0.4178
0.3765
**
MIDTENN
WESTTENN
0.3925
0.2043
0.4692
0.1462
0.3699
0.2123
0.4706
0.1529
FMRKT10
1.0645
1.0231
1.1370
0.9706
METRO
0.4194**
0.5385
0.4452
0.4882
*, **, *** denotes significance at the 10%, 5%, and 1% levels respectively based on t-tests.
a
For variable definitions see Table 1.
36
Table 3. Parameter Estimates from the Bivariate Probit Models for Estimating Factors Affecting
Awareness of Tennessee Farm Fresh and Pick Tennessee Products
Parameter Estimates for the Bivariate Probit Model
Awareness Equations
Independent Variablesa
Tennessee Farm Fresh
Pick Tennessee Products
**
-1.2527
1.1747*
Constant
(0.6220)
(0.6022)
-0.0044
-0.0215***
AGE
(0.0071)b
(0.0075)
0.2262
-0.7067*
SOMEHS
(0.3655)
(0.3668)
0.2113
-0.0934
HSGRAD
(0.2372)
(0.2359)
0.1382
-0.0937
SOMECOLL
(0.2626)
(0.2657)
-0.4827
-0.5676*
ASSOCDEG
(0.3179)
(0.3051)
0.0799
-0.0011
GRADDEG
(0.2383)
(0.2410)
-0.4180
-0.6153
PF_INCOME
(0.2012)
(0.2075)
0.0036
0.0060**
NEXTCNTY
(0.0025)
(0.0026)
0.0115**
0.0092*
INSTATE
(0.0051)
(0.0051)
-0.0001
-0.0102
INUS
(0.0098)
(0.0096)
0.0031
0.0198
OTHCNTRY
(0.0129)
(0.0172)
0.0143**
0.0039
VEGSIZE
(0.0067)
(0.0042)
0.0094**
0.0039
TOTDIRECTSALES
(0.0045)
(0.0041)
0.0050
0.0081
TOTINTERMSALES
(0.0053)
(0.0051)
-0.1037
-0.0727
EASTTENN
(0.1786)
(0.1842)
-0.1574
-0.2122
WESTTENN
(0.2332)
(0.2298)
-0.0729
-0.0870
FMRKT10
(0.0736)
(0.0733)
a
For variable definitions see Table 1.
b
Numbers in parenthesis are standard errors.
*, **, and *** represent statistical significance at 10%, 5%, and 1% levels, respectively.
37
Table 3. Continued
Independent Variablesa
METRO
EDUC_EVENTS
PUBLICATIONS
Likelihood value
Likelihood ratio
Correlation coefficient
Parameter Estimates for the Bivariate Probit Model
Awareness Equations
Tennessee Farm Fresh
Pick Tennessee Products
**
0.4121
0.1227
(0.1695)b
(0.1693)
0.1440***
0.1995***
(0.0464)
(0.0564)
***
0.6560
0.4588**
(0.1986)
(0.2060)
-290.5575
147.4400***
0.8278***
(0.0496)
a
For variable definitions see Table 1.
Numbers in parenthesis are standard errors.
*, **, and *** represent statistical significance at 10%, 5%, and 1% levels, respectively.
b
38
Table 4. Conditional Marginal Effects from the Bivariate Probit Model for Estimating Factors
Affecting Awareness of Tennessee Farm Fresh and P ick Tennessee Products
Marginal Effects of the Bivariate Probit Model
Prediction Conditions
AWARE_TFF=1 given
AWARE_PTP=1 given
Independent Variables
AWARE_PTP=1
AWARE_TFF=1
0.0037
-0.0044**
AGE
(0.0032)
(0.0018)
0.2710***
-0.3469**
SOMEHS
(0.0858)
(0.1609)
0.1280
-0.0636
HSGRAD
(0.0925)
(0.0620)
0.0934
-0.0505
SOMECOLL
(0.1066)
(0.0713)
-0.0812
-0.0698
ASSOCDEG
(0.1514)
(0.0967)
0.0396
-0.0130
GRADDEG
(0.0999)
(0.0528)
-0.0452
-0.0704*
PF_INCOME
(0.0834)
(0.0377)
0.0172
0.0246*
EDUC_EVENTS
(0.0219)
(0.0130)
**
0.1920
0.0059
PUBLICATIONS
(0.0742)
(0.0407)
0.0001
0.0009
NEXTCNTY
(0.0011)
(0.0006)
0.0032
0.0003
INSTATE
(0.0023)
(0.0011)
0.0028
-0.0024
INUS
(0.0043)
(0.0020)
-0.0039
0.0043
OTHCNTRY
(0.0072)
(0.0042)
0.0062**
-0.0014
VEGSIZE
(0.0031)
(0.0012)
0.0036*
-0.0006
TDS
(0.0020)
(0.0009)
0.0003
0.0011
TIN
(0.0024)
(0.0011)
-0.0322
-0.0006
EASTTENN
(0.0776)
(0.0385)
-0.0196
-0.0273
WESTTENN
(0.1025)
(0.0551)
*, **, and *** represent statistical significance at 10%, 5%, and 1% levels, respectively.
39
Table 4. Continued
Marginal Effects of the Bivariate Probit Model
Prediction Conditions
AWARE_TFF=1 given
AWARE_PTP=1 given
Independent Variables
AWARE_PTP=1
AWARE_TFF=1
-0.0126
-0.0091
FMRKT10
(0.0323)
(0.0161)
0.1763**
-0.0417
METRO
(0.0708)
(0.0361)
*, **, and *** represent statistical significance at 10%, 5%, and 1% levels, respectively.
40
50%
45%
40%
35%
Vegetable and melon farmers
Census 2007
30%
25%
20%
Fruit and tree nut farmers
Census 2007
15%
Sample data
10%
5%
0%
Under 25 25 to 34 35 to 44 45 to 54 55 to 64
65 and
over
Figure 1. Age distribution of sample data compared with the 2007 Census of Agriculture
41
Part 3: Factors Affecting Producer Participation in State Programs
Promoting Locally Grown Foods: The Case of Fruit and Vegetable Growers
in Tennessee
42
Abstract
U.S. governmental agencies have implemented a variety of programs to increase the
supply of Locally Grown Foods in response to growing popularity of these markets. Tennessee
currently has two state- funded programs promoting the consumption of Tennessee agricultural
products -Tennessee Farm Fresh and Pick Tennessee Products. Factors associated with produce
farmer participation in each of these programs are analyzed using a bivariate probit model.
Results suggest that participation in these programs was associated with use of Extension
resources, education, and fresh produce marketing. These results should help agencies attempting
to increase participation of producers in programs promoting locally grown foods.
Introduction
The sales of locally grown foods (LGF) in the U.S. reached$4.8 billion in 2008 and are
expected to grow to $7 billion by 2012 (Low and Vogel 2011; USDA/Agricultural Marketing
Services). While increased consumer interest in LGF may be new, the notion of states promoting
their own agricultural products has been around since at least the 1930’s (Patterson 2006) and a
number of states currently have programs designed to promote products grown in that state and
connecting producers with consumers seeking LGF. Previous studies have found that these
programs can increase consumer interest in, and sales of, specific products. Brooker and
Eastwood (1989) found that consumers in Knox County, Tennessee had positive attitudes
towards state logos. They noticed that consumer perceived state logos as useful tools to identify
LGF in metropolitan supermarkets. Govindasamy et al. (2004) evaluated the impact of the Jersey
Fresh program on cash receipts of farmers and the local economy in general. They estimated that
the $1.16 million investment in the Jersey Fresh campaign in 2000 generated $36.6 million in
revenues for New Jersey fruit and vegetable growers. Carpio and Isengildina-Massa (2010)
43
concluded that the South Carolina locally grown campaign increased consumers’ willingness to
pay for produce by 3.4%, it increased producer surplus by about three million dollars, and it
represented a return to investment of about 618% for the state of South Carolina.
In Tennessee, there are currently two state-funded programs to support and develop
markets for Tennessee-grown products. Pick Tennessee Products (PTP) was created by the
Tennessee Department of Agriculture (TDA) in 1986. In 2008, the Tennessee Department of
Agriculture, in cooperation with the Tennessee Farm Bureau, also created Tennessee Farm Fresh
(TFF). The purpose of these programs is to help farmers market their local products and inform
consumers about LGF markets.
Although previous studies have explored consumer awareness, perceptions and opinions
about state-sponsored marketing programs (Brooker and Eastwood 1989; Adelaja, Brumfield,
and Lininger 1990; Govindasamy et al. 1998b; Patterson et al. 1999; Govindasamy et al. 2004;
Carpio and Isengildina-Massa 2010; Onken and Bernard 2010; Onken, Bernard, and Pesek
2011), little research exists of producer response to state-sponsored promotion programs
(Govindasamy et al. 1998a). Govindasamy et al. (1998a) evaluated farmer awareness,
participation, perceptions and opinions about the Jersey Fresh logos (i.e., Jersey Fresh, quality
grading, and premium logos). About 93% of producers were aware of the Jersey Fresh Program,
and about 51% have used the Jersey Fresh Logos. Results suggest that producers who perceived
consumer high awareness of Jersey Fresh logos, who used other logos to identify fresh produce,
with more farming experience, and located in the agricultural zone of New Jersey were more
likely to use Jersey Fresh logos. In contrast, farmers with larger acreage and more than 75% of
their production being wholesaled were less likely to use these logos. Additionally, results
44
suggested that younger and more educated farmers were more likely to be willing to participate
in the Jersey Fresh Program in the future. They identified that the profile of farmers who willing
to use Jersey Fresh logos in the future was very similar to the one of farmers already using the
logos.
The goal of this study is to evaluate factors affecting fruit and vegetable producer
participation in TFF and PTP programs. The existence of two different programs within the same
state allows us to examine how observable differences in the two programs affect producer
participation. Information on the factors influencing producer participation in these programs can
help policy makers design and market similar programs in other states. The next section of this
second essay develops a theoretical model to explain producer participation in these programs.
The discussion of the data and methodology includes data description; theoretical model; and
methodology used to empirically analyze the factors influencing producer participation in statesponsored marketing programs in Tennessee. The results of this analysis are discussed next and
the final section concludes.
Data and Methodology
This study uses data from a survey of the entire population of fruit and vegetable
producers in Tennessee, as determined by the United States Department of Agriculture’s
National Agricultural Statistics Survey (USDA/NASS). The survey, a cover letter explaining the
importance of the survey, and a postage paid return envelope were mailed to Tennessee’s 1,954
fruit and vegetable producers on February 2, 2011. Reminder post cards were sent on February
24th . On March 24th , a second wave of surveys was sent to the producers who had not responded
45
to the initial mailing. Of the 1,954 surveys mailed, 587 were completed and returned for a
response rate of 30%.
The survey included an array of questions regarding outlets used to market produce, how
producers define a “local” market, barriers producers face when marketing their products, and
awareness and participation in the state-sponsored marketing programs. In addition, the survey
gathered information about farmer/farm business characteristics. More specifically, producers
were asked whether they were aware of each of Tennessee’s two programs – TFF and PTP.
Those responding that they were aware were then asked whether they participated in the program
or not. In addition to the survey data, secondary data such as the number of farmers markets per
county, and other county characteristics (e.g. metro/non- metro) that might be correlated with a
producer’s decision to participate in Tennessee’s programs promoting local foods were collected
from the United States Department of Agriculture’s Food Environmental Atlas
(http://www.ers.usda.gov/foodatlas, USDA 2011).
Theoretical Analysis
Fruit and vegetable producers are assumed to be rational decision makers who maximize
the discounted expected benefits from farming. Producers’ uncertainty about future income from
fruit and vegetable production may induce them to look for alternative marketing strategies to
improve benefits from farming. A producer decision to participate in a state-sponsored marketing
program can be seen as an attempt to boost benefits from farming through the potential increase
in sales, access to price premiums, and therefore contribution to local economies. Additionally,
producers may perceive the participation in these programs as an opportunity to contribute to the
wellness of their community, as they give access to local and maybe fresher foods (Govindasamy
46
et al. 1998a). Producers participating in state programs may develop a pride in participating in
these programs as they differentiate their produce using a logo that represents quality and
freshness. The utility producer receives from participating in state program m can be
represented by a random utility model such that:
(1)
U im   m' rim   im for m = 1, 2,
where rim is a vector of observed producer, farm, and region characteristics,  m is a vector of
unknown parameters associated with these variables, and  im is the error term. In this study two
programs will be evaluated: Tennessee Farm Fresh (m=1), and Pick Tennessee Produce (m=2).
A farmer will participate in a state-sponsored marketing program if the expected utility
*  0 ).
from participating is greater than zero ( U im

Note that U im
is an unobservable latent variable, but the decision to participate in a state
program is observable such that:
(2)


yim  1if U im  0 for m = 1,2,
0 if U im  0
where yim  1 if the producer decides to participate in state program m and yim  0 , otherwise.
This identity provides an empirically tractable approach to estimate the factors influencing the
participation in state-sponsored marketing programs.
47
Sample Selection
Since only those survey respondents who indicated that they were aware of TFF or PTP
were asked to indicate whether they participated in that program, participation is assumed to be
the result of a selection process that can be modeled as follows:
si1  1' xi1  ei1
si2   2' xi 2  ei 2
(3)
where si1  1 if a producer is aware of TFF, and zero otherwise; si2  1 if a producer is aware of
PTP, and zero otherwise; 1 and  2 are parameters associated with each awareness equation; ei1 ,
and ei2 , are random disturbances for each equation; and xi1 , and xi2 are vectors of observed
producer, farm, and area characteristics that may influence the likelihood that a producer is
aware of the two programs.
A Heckman selection probit model is used to examine the factors affecting producer
participation in TFF and PTP, subject to awareness to these programs. Assuming that the error
terms in the selection equations presented in (3) are normally distributed with μ m  0 , and
Var (eim )  σ m2 , it can be shown that:
(4)
PAm  Pr(sim  1)  1  (( m' xim )) m= 1,2,
where PAm is the probability of a farmer being aware of a state-sponsored marketing program m,
and Φ represents the standard normal cumulative distribution function. The symmetric qualities
of the standard normal distribution function can be used to show that:
48
1  Φ(( β xi ))  Φ( β xi ) .
(5)
Therefore, the probability of a farmer being aware of a state program can be represented as:
PAm  Φ( βm' xim ) .
(6)
Given equations (4) and (6), the sample likelihood function can be written as:
L   sim 1(m' xim ) sim 0((m' xim ))
(7)
Assuming the error term in equation (1) is distributed standard normal with  ym  0 and
Var( i )   ym 2 , then
PPm  Pr( yim  1| sim  1)  Pr(U m*  0 | sim  1)
(8)
 Pr(εim  (α'm rim )|Φ( βm' xim ))
 Φ(α'm rim )|Φ( βm' xim ) ,
where PPm is the probability of a producer participating in state program m given awareness of
this program. The sample likelihood function can then be written as:
(9)
L   ymi 1|smi 1(( m' rim ),  ) ymi 0|smi 1(( m' rim ),   ) ,
where ρ is the correlation between the error term in the selection equation, e im , and the error
term in the outcome equation,  im . This correlation measures the level of association between the
unobserved determinants of awareness of TFF and PTP and the unobserved determinants of
participation in these programs and it is a result of the non-random nature of the samples used to
49
evaluate participation in TFF and PTP. If ρ  0 , the Heckman correction is appropriate to
estimate the parameters in equation (1). If ρ  0 , the model has failed to identify any selection
bias and a probit model can be used to estimate participation in each program.
However, there may be unobserved determinants of participation in TFF and PTP that are
correlated. If so, the error terms of the participation equations will be correlated. The TFF and
PTP programs are similar in nature. Both TFF and PTP were created in order to link producers of
LGF with consumers seeking LGF. In addition, both programs offer many of the same services
including listing on an on-line directory and the right to use program logos. Thus, it seems likely
that many of the same factors that influence participation in one program will influence
participation in the other. If the error terms of the participation equation are found to be
correlated a bivariate probit model is appropriate to estimate the equations described in (1)
(Greene 2003; Christofides, Stengos, and Swidinsky 1997).
Empirical Model
Awareness is a necessary condition for participation in TFF and PTP, therefore the
equation for awareness is set to be the selection equation in the Heckman selection probit
estimation. Descriptions of the variables used in the selection equations are presented in Tables 1
and 2.
Producer characteristics hypothesized to affect aware ness are: age (AGE); educational
attainment, represented by a set of dummy variables for some high school (SOMEHS), high
school graduate (HSGRAD), some college (SOMECOLL), an associate’s degree (ASSOCDEG),
a bachelor’s degree (BACHDEG), and a graduate degree (GRADDEG); the percentage of
household income from farming, represented by a single dichotomous variable for less than 25
50
percent (PF_INCOME); the number of University/Extension educational events or presentations
involving produce marketing strategies attended in the past five years (EDUC_EVENTS); and a
dichotomous variable indicating whether respondent has used any University/Extension
educational publications regarding produce marketing strategies in the past five years
(PUBLICATIONS).
Farm enterprise characteristics included in the awareness equation are: size of the fruit
and vegetable operation in acres (VEGSIZE); percentage of fruit and vegetable sales made
directly to consumers (TDS), intermediaries (TIN), and retail outlets (TRE); and the percent age
of direct to consumer sales made to consumers in the producer’s own county (YOURCNTY),
neighboring counties (NEXTYCNTY), elsewhere in the state of Tennessee (INSTATE),
elsewhere in the U.S. (INUS), and elsewhere in other countries (OTHCNTRY).
The characteristics of the region in which the producer operates that are included in the
awareness equation are: geographic location, represented by dichotomous variables for East
Tennessee (EASTTENN), Middle Tennessee (MIDTENN), and West Tennessee (WESTTENN);
a dichotomous variable for whether the producer operates in a metropolitan county (METRO);
and the number of farmers markets located within the producer’s county of operation (FMRKT).
Hypotheses about the impact (positive or negative) of all the variables described above on the
awareness of TFF and PTP are described in the first essay of this thesis.
Descriptions of the variables used in the participation equations are presented in Table 3.
Producer characteristics hypothesized to affect participation are: age (AGE); whether the
producer has attained a bachelor or graduate degree (BACH_GRAD); the percentage of
51
household income coming from farming, expressed by a dichotomous variable for less than 25
percent (PF_INCOME); extent to which the producer agrees that labeling his or her produce as
locally grown will increase sales on a scale of one to five where one represents ”strongly
disagree” and five represents ”strongly agree” (BENEFIT_SALES); number of
University/Extension events related to produce marketing strategies attended in the past five
years (EDUC_EVENTS); and whether the producer used University/Extension publications to
obtain information on produce marketing strategies in the past five years (PUBLICATIONS).
Age has been found to have a negative influence on use of state- logo programs by
farmers (Govindasamy et al. 1998a). Thus, younger farmers are expected to be more likely to
participate in state-sponsored marketing programs. Education is expected to be positively
correlated with participation in these programs. Prior research found that producers with more
than college education were more likely to participate in state-sponsored marketing programs in
the future (Govindasamy et al. 1998a). Additionally, previous studies found producers
marketing their produce directly to consumers to be more educated than those not using these
outlets, perhaps because these marketing outlets may require additional skills or abilities beyond
those needed for the management of an agricultural operation (Uva 2002; Hunt 2007; Uematsu
and Mishra 2011). Hence, given that producers participating in state-sponsored marketing
programs are more likely to use direct-to-consumer outlets (Govindasamy et al.1998), they are
also expected to be more educated than those not participating in these programs.
The percentage of household income coming from farming is hypothesized to have a
positive impact on participation as households with a greater percentage of income from farming
are likely to be willing and able to devote more time to implementing new marketing strategies.
52
Marketing local foods may be more time- intensive (e.g. more time to contact buyers) than other
strategies, and therefore a larger percentage of income from farming may imply a higher
probability to participate in programs promoting local foods (D’Souza, Cyphers, and Phipps
1993).Producer’s believes that labeling produce as local will increase produce sales is expected
to be positively correlated with participation. Given that state-sponsored marketing programs are
meant to increase popularity of products grown in the state and potentially increase sales of these
products, it is expected that producers perceiving the effectiveness of these programs in attaining
this goal will increase producer likelihood to participate (Govindasamy et al. 1998a).
Attendance at University/Extension educational events and use of University/Extension
publications are hypothesized to increase participation. Information plays a key role in the
adoption of new management practices including marketing and Extension services can be an
effective tool in delivering the information needed for farmers to make informed decisions about
new marketing strategies (Nowak 1987; Knowler and Bradshaw 2007).
Characteristics of the producer’s farming operation included in the outcome equation are:
size of the producer’s fruit and vegetable operation in acres (VEGSIZE); percentage of sales
made directly to consumers (TDS); and whether the operation markets fresh fruits and vegetables
(FRESH).
While some studies suggest that that the number of acres being farmed has a negative
impact on participation in state-sponsored marketing programs (Govindasamy et al. 1998a),
other results suggest that adoption of new marketing strategies does not seem to have a scale
effect where the “cost” of adoption of a new marketing strategy is spread over the number of
53
acres farmed (D’Souza, Cyphers, and Phipps 1993). Thus, there is no a priori hypothesis with
respect to participation and size of operation. On the other hand, the percentage of sales made
directly to consumers is hypothesized to be positively correlated with participation. Previous
research found that producers categorized as primary wholesalers (i.e., more than 75 percent of
their production being sold in the wholesale market) were less likely to participate in these
programs (Govindasamy et al. 1998a). Therefore, we expected that those who primary sell
through direct-to-consumer outlets as oppose to wholesale markets may be more likely to
participate in state-sponsored marketing programs. The marketing of fresh fruits and vegetables
is hypothesized to increase participation solely in the TFF program as TFF focuses on the
promotion of fresh products grown in Tennessee.
Characteristics of the area in which the prod ucer operates included in the participation
equations are: region of Tennessee producer is located in, expressed by dichotomous variables
for East (EASTTENN), Middle (MIDTENN), and West (WESTTENN); and whether producer is
located in a metropolitan county (METRO).
Producers located in metropolitan counties are expected to be more likely to participate in
TFF and PTP. Producers living in metropolitan areas rely heavily on urban markets. More than
half of the farms with local food sales are situated in metropolitan counties (Low and Vogel
2011).
54
Results and Discussion
Sample Overview and Descriptive Statistics
After eliminating observations with missing data, 301 and 302 responses were suitable
for the analysis of awareness of the TFF and PTP programs, respectively. For the joint analysis
of participation in TFF and PTP 104 responses were suitable for analysis after eliminating
observations with missing data. The survey design made individuals who were not aware of TFF
or PTP not to answer the question of participation, and this explains in part why for the
participation in TFF and PTP analysis only 104 observations were used. The average age of
respondents included in the analysis of factors affecting participation in TFF and PTP (n=104)
was 58 years which is the average age of farmers in Tennessee according to the 2007 Census of
Agriculture (USDA/NASS). The age distribution from respondents follows closely age
distribution among vegetable and melon farmers, and fruit and nut farmers in Tennessee (Figure
2). The proportions of farmers in each age category are similar when comparing respondents age
and age of vegetable and melon farmers in Tennessee. The sample used in this study had a larger
proportion of farmers in the 35 to 44, 45 to 54 and 55 to 64 age categories when compared to
Tennessee fruit and nut farmers. However, the proportion of Tennessee fruit and nut farmers in
the 65 and over category was larger compared to the sample respondents. About 54% of
respondents have attained either a bachelor’s degree or graduate degree. About 52% of
respondents reported that less than 25% of their household income came from farming. On
average, respondents somewhat agreed with the statement that labeling produce as “locally
grown” would increase “the number of customers willing to buy my produce.” Respondents
attended an average of 2.2 University/Extension educational events relating to marketing
strategies for selling produce in the past five years. About 52% of respondents have used
55
University/Extension publications to obtain information about how to better market their produce
in the past five years.
The average size of respondents’ fruit and vegetable operations is 12 acres. About 87% of
respondent fruit and vegetable sales were made through direct-to-consumer outlets. About 92%
of producers indicated that they sell fresh fruits and vegetables, which is very close to the
percentage of producers selling fresh vegetables in Tennessee (91%) according to the 2007
Census of Agriculture (USDA/NASS). Approximately 37% of respondents were located in East,
52% in Middle, and 12% in West Tennessee. About 52% of the respondents’ fruit and vegetable
operations are located in metropolitan counties.
Approximately 23% of the respondents participate in the TFF program and 39%
participate in the PTP program. The difference in the percentage of respondents participating in
each program may be due to the fact that PTP was established in 1986 and TFF is still relatively
new, having been established in 2008. The difference could also be attributed to the fact that
participation in TFF entails an annual fee ($100.00) while PTP does not, although the TFF fee
guarantees an array of marketing tools and benefits (i.e. TFF logo, TFF stickers, price cards, TFF
reusable bags, and free access to workshops offered through the University of Tennessee Center
for Profitable Agriculture).
Mean values of producer, operation, and area characteristics for respondents who
participate in the programs and for those who do not are reported in Table 4. Differe nces in these
mean values between those who participate and those who do not are evaluated using t-tests. On
average, participants in both programs have larger fruit and vegetable operations, have attended
more University/Extension educational events related to fruit and vegetable marketing in the last
56
five years, and were more likely to have used University/Extension publications to obtain
information on how to better market their produce in the last five years. A larger proportion of
respondents participating in PTP had attained a bachelor or graduate degree (66%) than those not
participating in the program (46%). The proportion of respondents with 25% or less of their
income coming from farming was significantly lower among PTP participants than among no nparticipants. This finding supports the hypothesis that respondents with higher percentages of
income coming from farming are more likely to participate in these programs to improve
marketing strategies because they rely more heavily on farm incomes (D’Souza, Cyphers, and
Phipps 1993).
A greater proportion of respondents not participating in TFF sell fresh fruits and
vegetables (95%) compared to respondents participating in this program (83%). A significantly
higher proportion of respondents whose operation was located in East Tennessee were not
participating in PTP than were participating, with 43% and 27%, respectively. Finally, a
significantly larger proportion of respondents operating in metropolitan counties are participating
in PTP (63%) than those who were not participating (44%). This result is consistent with the
hypothesis that because urban markets are used by a large number of producers marketing their
products as local foods producers located near to these markets are more likely to participa te in
state-sponsored marketing programs (Low and Vogel 2011).
Evaluation of Factors Affecting Conditional and Joint Probabilities of Participation in TFF and
PTP: A Bivariate Probit Model
Heckman selection probit models were estimated to identify factors affecting
participation in TFF and PTP. However, the correlation coefficients between the selection
57
equation and the outcome equation were not statistically significant for either the TFF or PTP
models, indicating that probit models could be used for this analysis (Table 5). However, it is
hypothesized that because of similarities in the TFF and PTP programs, there may be unobserved
determinants that are likely to have an analogous influence on participation in these programs,
and, as a result, the error terms of the two participation equations are likely to be correlated.
Therefore, a bivariate probit model was used. The correlation coefficient between the residuals
for the participation in TFF and PTP equations was positive and significant at the 1% leve l. This
result indicates that the error terms of the TFF and PTP participation equations are correlated and
therefore, the use of the bivariate probit model is appropriate. A likelihood ration test also
indicated the model was overall significant at the 1% level. Results from the bivariate probit
model can be seen in Table 5. As explained above, only respondents who reported to be aware of
TFF or PTP were asked to answer the question whether they participate or not on these
programs. Therefore, only individuals aware of both programs were considered in the bivariate
probit model but no corrections were made for potential selection bias given the results obtained
from the Heckman selection probit.
Marginal effects of the various explanatory variables consid ered on the conditional and
joint probabilities of participation in these two programs are presented in Table 6. For this
analysis we only present the marginal effects for one conditional probability and three joint
probabilities because these were the ones considered relevant for this study3 . Marginal effects
were estimated for the probability of participating in TFF given participation in PTP,
Pr( yi1  1| yi 2  1| ri1 , ri 2 ) , probability of jointly participating in both programs,
3
Other marginal effects are availab le fro m author upon request.
58
Pr( yi1  1, yi 2  1| ri1 , ri 2 ) , and probability of participating in PTP but not in TFF,
Pr( yi1  0, yi 2  1| ri1 , ri 2 ) . Conditional probabilities are related with the probability of an event
given prior information (e.g. probability of participation in TFF given prior information that the
respondent is participating in PTP). Joint probabilities deal with events that happened
simultaneously without any prior information. Given that information about the decision making
process behind participation in TFF and PTP is limited, both, conditional and joint probabilities
were considered. Estimation of marginal effects followed Christofides, Stengos, and Swidinsky
(1997) approach. The number of University/Extension educational events attended in the past
five years (EDUC_EVENTS) and whether the producer sold fresh fruits and vegetables
(FRESH) had a statistically significant effect on the probability of participating in TFF given that
the producer was already participating in PTP. Attendance at an additional University/Extension
educational event increased the likelihood of producer participation in TFF by 5%, given that the
producer was already participating in PTP. If the producer was selling fresh fruits and vegetables
(FRESH) he/she was 48% less likely to participate in TFF, given participation in PTP. This
result indicates that, even though a main focus of TFF is to promote locally grown fresh produce,
fruit and vegetable producers are less likely to participate in this program if they are already
participating in PTP. Contrary to the hypothesis that producers selling fresh produce are more
likely to participate in TFF given that the focus of this program is the promotion of fresh produce
grown in Tennessee, for this sample, it seems that a larger proportion of operations selling fresh
produce are not participating in TFF. These results may be explained by the fact that PTP
promotes all products grown in Tennessee not only fresh produce, and therefore given its
coverage it may be perceived by producers as a program having a larger impact on consumer
59
preferences. Therefore, a produce farmer may be less interested in participating in alternative
programs focusing on fresh produce only. Additionally, the fact that PTP has been around for
about twenty four years may increase producer confidence in the effectiveness of the campaign
reducing the need for participation in any other program specializing in the type of products they
are marketing (i.e. fresh produce).
Producers who have either a bachelor or graduate degree (BACH_GRAD) or who sell
fresh fruits and vegetables (FRESH) are more likely to participate in PTP and not in TFF.
Producers who have either a bachelor’s or graduate degree are 21% more likely to participate in
PTP and not participate in TFF, and producers who sell fresh fruits and vegetables are 19% more
likely to participate in PTP and not in TFF. As hypothesized, producers who have attained higher
levels of education are more likely to participate in program promoting local foods as the use of
marketing outlets associated with LGF may require additional skills or abilities not related with
the management of an agricultural operation (Uva 2002; Hunt 2007; Uematsu and Mishra 2011).
The size of fruit and vegetable operation (VEGSIZE), the number of
University/Extension educational events attended in the past five years (EDUC_EVENTS), and
the use of University/Extension publications in the past five years (PUBLICATIONS) had a
positive impact on the joint probability of participating in both TFF and PTP. Although
statistically significant, the effect of fruit and vegetable operation size on the likelihood of
participation in both programs was very small in magnitude. Attendance to University/Extension
events increase the likelihood of jointly participating in TFF and PTP by 4% and the use of
University/Extension publications increased the likelihood of jointly participation in both
programs by 14%. The correlation between use of University/Extension resources, such as
60
educational events and publications, and program participation is expected given that Extension
services are an important source of information for farmers looking for new farming practices
and marketing strategies (Nowak 1987; Knowler and Bradshaw 2007).Whether the producer sold
fresh fruits and vegetables decreased the likelihood of participation in TFF and PTP by 19%.
Conclusions
Federal and state agencies have implemented a variety of programs to increase the supply
of LGF in response to growing popularity of these markets among consumers. Tennessee
currently has two state- funded programs in place to support and develop markets for Tennesseegrown products: Tennessee Farm Fresh (TFF) and Pick Tennessee Products (PTP). A bivariate
probit regression was used to evaluate the impact of characteristics of the producer, the
producer’s operation, and the county in which the producer’s operation is located on producer
participation in TFF and PTP.
Different factors affected the likelihood of participating in one program given the
participation in the alternative program, and the likelihood of jointly participating in both
programs (i.e., TFF and PTP). Attendance to University/Extension educational events and
whether the producer sold fresh fruits and vegetables significantly affected participation in TFF,
given producer participation in PTP. On the other hand, higher levels of education (i.e.
bachelor’s or graduate degree) and whether the producer sold fresh fruits and vegetables
significantly affected participation in PTP, given producer participation in TFF. S ize of fruit and
vegetable operation, attendance to University/Extension educational events, and the use of
University/Extension publications had a significant effect on the joint probability of participation
61
in both TFF and PTP. It is important to notice that if a producers is selling fresh produce he/she
is more likely to participate in PTP, given participation in TFF, while a producer selling fresh
fruits and vegetables is less likely to participate in TFF, given he/she is already participating in
PTP. This result is surprising given that the focus of TFF is promotion of fresh products grown in
Tennessee. This result may suggest further exploration of the differences in the profile of
producer who could potentially participate in TFF and PTP, as well as their needs, given that the
marketing of fresh produce does not necessarily increase interest in participating in TFF if a
producer is already participating in PTP. Additionally, it would be important to identify the
profile of those producers who may benefit from participation in both programs simultaneously,
so that policy makers can better satisfy the needs of this particular clientele.
Additionally, results point at the importance of Extension as a source of information for
producers who may be interested in participating in programs promoting LGF. Attendance to
University/Extension events and the use of University/Extension publications were significant
factors affecting participation in TFF and PTP. Therefore, policy makers should continue
collaborating with Extension services to increase participation in marketing programs but may
also explored alternative information sources for those who do not use Extension as their main
source of information when looking for alternative marketing strategies.
62
References
Adelaja, A.O., R.G. Brumfield, and K. Lininger. 1990. Producer Differentiation and State
Promotion of Farm Produce: An Analysis of the Jersey Fresh Tomato. Journal of Food
Distribution Research 21(3): 73-85.
Brooker, J.R., and D.B. Eastwood. 1989. Using State Logos to Increase Purcha se of Selected
Food Products. Journal of Food Distribution Research 20(1): 175-183.
Carpio, C. and O. Isengildina-Massa. 2009. “Measuring the Potential Economic Impact of a
Regional Agricultural Promotion Campaign: The Case of South Carolina.” Paper
presented at the Southern Agricultural Economics Association Annual Meeting, Atlanta,
Georgia, January 31-February 3, 2009.
Christofides, L.N., T. Stengos, and R. Swidinsky. “On the calculation of marginal effects in the
bivariate probit model.” Economic Letters 54(1997): 203-208.
D’Souza, G., D. Cyphers, and T. Phipps. “Factors Affecting the Adoption of Sustainable
Agricultural Practices.” Agricultural Resource Economics Review 40(1993): 159-165.
Govindasamy, R., A. Pingali, J. Italia, and D. Thatch. 1998a. Producer Response to StateSponsored Marketing Programs: The Case of Jersey Fresh. Department of Agricultural,
Food and Resource Economics Rutgers University, P Series 02137-3-98.
Govindasamy, R., B. Schilling, K. Sullivan, C. Turvey, L. Brown, and V. P uduri. 2004. Returns
to the Jersey Fresh Promotional Program: The Impacts of the Promotional Expenditures
on Farm Cash Receipts in New Jersey. Working paper, Department of Agricultural, Food
and Resource Economics, and the Food Policy Institute, Rutgers, The State University of
New Jersey.
Govindasamy, R., J. Italia and D .Thatch. 1998b. "Consumer Awareness of State-Sponsored
Marketing Programs: An Evaluation of the Jersey Fresh Program." Journal of Food
Distribution Research. 29(1998 d): 7-15.
Greene, W.H. Econometric Analysis. 5th Edition. Upper Saddle, NJ: Prentice Hall, 2003.
Hunt, R. “Consumer Interactions and Influences on Farmers’ Market Vendors.” Renewable
Agriculture and Food Systems 22(2007): 54-66.
Knowler, D., and B. Bradshaw. “Farmer’s adop tion of conservation agriculture: A review and
synthesis of recent research.” Food Policy 32(2007): 25-48.
Low, S.A., and Vogel, S. Direct and Intermediated Marketing of Local Foods in the United
States. Washington DC: U.S. Department of Agriculture, ERS Economic Research Rep.
128, November, 2011.
63
Nowak, P. J.“The adoption of agricultural conservation technologies: economic and diffusion
explanations.” Rural Sociology 52(1987): 208–220.
Onken, K.A., and J.C. Bernard 2010. Catching the “Local” Bug: A Look at State Agricultural
Marketing Programs. Choices 25(1). Internet site: http://www.choicesmagazine.org/
magazine/article.php?article=112 (Accessed March 8, 2012).
Onken, K.A., J.C. Bernard, and J.D. Pesek Jr. 2011. Comparing Willingness to Pay for Organic,
Natural, Locally Grown, and State Marketing Program Promoted Foods in the MidAtlantic Region. Agricultural and Resource Economics Review 40 (1): 33-47.
Patterson, P.M. “State-Grown Promotion Programs: Fresher, Better?” Choices 21(2006): 41-46.
Patterson, P.M., H. Oloffson, T. J. Richards, and S. Sass. 1999. An Empirical Analysis of State
Agricultural Product Promotions: A Case Study of Arizona Grown. Agribusiness: An
International Journal 15: 179-196.
U.S. Department of Agriculture, National Agricultural Statistics Service (USDA/NASS). 2007
Census of Agriculture. Internet site: http://www.agcensus.usda.gov/Publications/2007/
index.asp (Accessed October 10 , 2011).
Uematsu, H., and A. K. Mishra.“Use of Direct Marketing Strategies by Farmers and Their
Impact on Farm Business Income.” Agricultural and Resource Economics Review
40(2011): 1 – 19.
Uva, W-F.L. “An Analysis of Vegetable Farms’ Direct Marketing Activities in New York State.”
Journal of Food Distribution Research 33(2002): 186–189.
64
Appendix
Table 5. Definitions and Descriptive Statistics of Variables Affecting Awareness of Tennessee
Farm Fresh (n=301)
Variable
Description
Mean
C. Dependent
Variables
=1 if farmer is aware of Tennessee Farm Fresh,
AWARE_TFF
0.3787
zero otherwise
D. Independent
Variables
AGE
Age of producer in years
60.6013
=1 if some high school is the highest level of
SOMEHS
0.0631
education attained by the farmer, zero otherwise
=1 if high school diploma is the highest level of
HSGRAD
0.2060
education attained by the farmer, zero otherwise
=1 if some college is the highest level of education
SOMECOLL
0.1462
attained by the farmer, zero otherwise
=1 if an associate’s degree is the highest level of
ASSOCDEG
0.0963
education attained by the farmer, zero otherwise
=1 if a bachelor’s degree is the highest level of
BACHDEG
0.2658
education attained by the farmer, zero otherwise
=1 if a graduate degree is the highest level of
GRADDEG
0.2226
education attained by the farmer, zero otherwise
=1 if less than 25% of farmer household income
PF_INCOME
0.7043
comes from farming
The number of educational events the farmer has
EDUC_EVENTS
1.1296
attended in the past 5 years
=1 if the farmer has used University/Extension
PUBLICATIONS
0.2990
publications in the past 5 years
Percent of direct sales to consumers in the county
YOURCNTY
68.8239
where the farmer operates
Percent of direct sales to consumers in neighboring
NEXTCNTY
23.6678
counties of where the farmer operates
Percent of direct sales to consumers elsewhere in
INSTATE
5.0681
the state
Percent of direct sales to consumers elsewhere in
INUS
1.8090
the country
Percent of direct sales to consumers in other
OTHCNTRY
0.6312
countries
VEGSIZE
Size of fruit and vegetable operation in acres
8.9132
Percent of direct sales obtained from direct to
TOTDIRECTSALES
85.1728
consumer outlets
65
Table 5. Continued
Variable
TOTINTERMSALES
TOTRETAILSALES
EASTTENN
MIDTENN
WESTTENN
FMRKT
METRO
Description
Percent of direct sales obtained from direct to
intermediary outlets
Percent of direct sales obtained from direct to retail
outlets
=1 if the farmer is located in East Tennessee, zero
otherwise
=1 if the farmer is located in Middle Tennessee,
zero otherwise
=1 if the farmer is located in West Tennessee, zero
otherwise
The number of farmer’s markets in the county
where the farmer operates
=1 if farmer is located in a metropolitan county,
zero otherwise
66
Mean
7.5914
7.2359
0.3987
0.4319
0.1694
1.0664
0.4651
Table 6. Definitions and Descriptive Statistics of Variables Affecting Awareness of Pick
Tennessee Products (n=302)
Variable
Description
Mean
E. Dependent
Variables
=1 if farmer is aware of Pick Tennessee Products,
AWARE_PTP
0.5099
zero otherwise
F. Independent
Variables
AGE
Age of producer in years
60.5927
=1 if some high school is the highest level of
SOMEHS
0.0662
education attained by the farmer, zero otherwise
=1 if high school diploma is the highest level of
HSGRAD
0.2086
education attained by the farmer, zero otherwise
=1 if some college is the highest level of education
SOMECOLL
0.1457
attained by the farmer, zero otherwise
=1 if an associate’s degree is the highest level of
ASSOCDEG
0.0960
education attained by the farmer, zero otherwise
=1 if a bachelor’s degree is the highest level of
BACHDEG
0.2550
education attained by the farmer, zero otherwise
=1 if a graduate degree is the highest level of
GRADDEG
0.2285
education attained by the farmer, zero otherwise
=1 if less than 25% of farmer household income
PF_INCOME
0.6921
comes from farming
The number of educational events the farmer has
EDUC_EVENTS
1.0993
attended in the past 5 years
=1 if the farmer has used University/Extension
PUBLICATIONS
0.2980
publications in the past 5 years
Percent of direct sales to consumers in the county
YOURCNTY
69.0530
where the farmer operates
Percent of direct sales to consumers in neighboring
NEXTCNTY
23.5960
counties of where the farmer operates
Percent of direct sales to consumers elsewhere in
INSTATE
5.0513
the state
Percent of direct sales to consumers elsewhere in
INUS
1.6705
the country
Percent of direct sales to consumers in other
OTHCNTRY
0.6291
countries
VEGSIZE
Size of fruit and vegetable operation in acres
9.5111
Percent of direct sales obtained from direct to
TOTDIRECTSALES
85.3378
consumer outlets
Percent of direct sales obtained from direct to
TOTINTERMSALES
7.4338
intermediary outlets
67
Table 6. Continued
Variable
TOTRETAILSALES
EASTTENN
MIDTENN
WESTTENN
FMRKT
METRO
Description
Percent of direct sales obtained from direct to retail
outlets
=1 if the farmer is located in East Tennessee, zero
otherwise
=1 if the farmer is located in Middle Tennessee,
zero otherwise
=1 if the farmer is located in West Tennessee, zero
otherwise
The number of farmer’s markets in the county
where the farmer operates
=1 if farmer is located in a metropolitan county,
zero otherwise
68
Mean
7.2285
0.3974
0.4272
0.1755
1.0861
0.4636
Table 7. Definitions and Descriptive Statistics of Variables Affecting Participation in TFF and
PTP (n=104)
Variable
Description
Mean
G. Dependent
Variables
=1 if farmer is participating in Tennessee Farm
PART_TFF
0.2308
Fresh, zero otherwise
=1 if farmer is participating in Pick Tennessee
PART_PTP
0.3942
Products, zero otherwise
H. Independent
Variables
AGE
Age of producer in years
58.0481
=1 if producer has attained a bachelor’s or graduate
BACH_GRAD
0.5385
degree
=1 if less than 25% of farmer household income
PF_INCOME
0.5192
comes from farming
VEGSIZE
Size of fruit and vegetable operation in acres
11.7740
Percent of direct sales obtained from direct to
TDS
87.2115
consumer outlets
=1 if the farmer is located in East Tennessee, zero
EASTTENN
0.3654
otherwise
=1 if the farmer is located in Middle Tennessee,
MIDTENN
0.5192
zero otherwise
=1 if the farmer is located in West Tennessee, zero
WESTTENN
0.1154
otherwise
=1,…,5 based on degree to which farmers believe
BENEFIT_SALESa labeling their produce as “locally grown” will
4.1250
increase sales
=1 if farmer is located in a metropolitan county,
METRO
0.5192
zero otherwise
The number of educational events the farmer has
EDUC_EVENTS
2.2019
attended in the past 5 years
=1 if the farmer has used University/Extension
PUBLICATIONS
0.5192
publications in the past 5 years
FRESH
=1 if the farmer sells fresh fruits and vegetables
0.9231
a
Responses were 1= “strongly disagree”, 2= “somewhat disagree”, 3= “neither agree nor
disagree”, 4= “somewhat agree”, and 5= “strongly agree.”
69
Table 8. Variable Means for Respondents Participating in the Tennessee Farm Fresh, Pick
Tennessee Products, and Both Programs and Those Not Participating
Pick Tennessee
Tennessee Farm Fresh
Both Programs
Products
Do Not
Participate
Do Not
Participate
Do Not
Participate
Independent
Participate
(n=24)
Participate
(n=41)
Participate
(n=21)
Variablesa
(n=80)
(n=63)
(n=83)
AGE
57.9375
58.4168
58.9206
56.70732
58.0602
58.0000
BACH_GRAD
0.5250
0.5833
0.4603
0.6585**
0.5181
0.6190
**
*
PF_INCOME
0.5625
0.3750
0.6032
0.3902
0.5663
0.3333
VEGSIZE
8.8250
21.6042***
8.6905
16.5122**
8.7530
23.7142***
TOTDIRECTSALES 89.0250
81.1667
88.9048
84.6098
89.0482
79.9524
*
EASTTENN
0.4000
0.2500
0.4286
0.2683
0.3855
0.2857
MIDTENN
0.4875
0.6250
0.4921
0.5610
0.5060
0.5714
WESTTENN
0.1125
0.1250
0.0794
0.1707
0.1084
0.1429
BENEFIT_SALES
4.1000
4.2083
4.0000
4.3171
4.0843
4.2857
*
METRO
0.5000
0.5833
0.4444
0.6341
0.4940
0.6190
EDUC_EVENTS
1.4375
4.7500***
1.2540
3.6585***
1.5181
4.9048***
PUBLICATIONS
0.4250
0.8333***
0.3968
0.7073***
0.4337
0.8571***
FRESH
0.9500*
0.8333
0.9365
0.9024
0.9277
0.9048
*, **, *** denotes significance at the 10%, 5%, and 1% levels respectively based on t-tests.
a
For variable definitions see Table 1.
70
Table 9. Heckman Sample Selection Model Estimation of Participation in the Tennessee Farm
Fresh and Pick Tennessee Products Program Given Awareness of These Programs
Dependent Variable
Aware of TFF Participating in Aware of PTP Participating in
(n=114)
TFF (n=28)
(n=154)
PTP (n=58)
Independent Variables
Coefficient
Coefficient
-2.0040***
-2.7362**
0.9923
-1.6973**
Constant
(0.7502)
(1.2595)
(0.6387)
(0.8108)
-0.0019
0.0115
-0.0200***
0.0026
AGE
a
(0.0074)
(0.0150)
(0.0075)
(0.0108)
0.2836
-0.6259*
SOMEHS
(0.3796)
(0.3728)
-0.0021
-0.1674
HSGRAD
(0.2577)
(0.2570)
0.0410
-0.1871
SOMECOLL
(0.2755)
(0.2780)
-0.5343
-0.5821*
ASSOCDEG
(0.3320)
(0.3153)
0.1372
0.0175
GRADDEG
(0.2427)
(0.2446)
-0.0468
0.3677
BACH_GRAD
(0.3095)
(0.2493)
-0.4038*
-0.0813
-0.6468***
-0.4615
PF_INCOME
(0.2196)
(0.3191)
(0.2150)
(0.2850)
0.0054**
0.0073**
NEXTCNTY
(0.0027)
(0.0028)
0.0127**
0.0089
INSTATE
(0.0056)
(0.0055)
0.0002
-0.0109
INUS
(0.0112)
(0.0108)
0.0023
0.0252
OTHCNTRY
(0.0139)
(0.0243)
0.0201**
0.0295***
0.0040
0.0097
VEGSIZE
(0.0082)
(0.0093)
(0.0068)
(0.0076)
0.0145**
-0.0047
0.0048
-0.0008
TDS
(0.0057)
(0.0065)
(0.0045)
(0.0039)
0.0072
0.0089
TIN
(0.0064)
(0.0057)
0.1069
0.1205
BENEFIT_SALES
(0.1615)
(0.1069)
-0.7970**
FRESH
(0.4446)
*, **, *** denotes significance at the 10%, 5%, and 1% levels respectively.
a
Standard errors are in parenthesis.
71
Table 9. Continued
Independent Variables
EASTTENN
WESTTENN
FMRKT
METRO
EDUC_EVENTS
PUBLICATIONS
ρ
Dependent Variable
Aware of TFF Participating in Aware of PTP Participating in
(n=114)
TFF (n=28)
(n=154)
PTP (n=58)
Coefficient
Coefficient
-0.1041
0.0375
-0.1348
-0.1335
(0.1849)a
(0.3399)
(0.1872)
(0.2551)
-0.4444*
0.0101
-0.3623
0.7097*
(0.2651)
(0.4516)
(0.2408)
(0.4042)
-0.0275
-0.0777
(0.0765)
(0.0741)
0.3869**
0.1928
0.0836
0.1876
0.1757)
(0.2971)
(0.1753)
(0.2504)
0.1279***
0.1950***
0.2009***
0.1496***
0.0460)
(0.0533)
(0.0615)
(0.0546)
***
***
***
0.6843
1.0208
0.4880
0.3636
0.2099)
(0.3712)
(0.2149)
(0.2715)
0.9989
0.2806
 statistic (H : ρ  0)
0.3204
0.6065
0
Log likelihood value
-191.4954
-238.4091
*, **, *** denotes significance at the 10%, 5%, and 1% levels respectively.
a
Standard errors are in parenthesis.
2
72
Table 10. Parameter Estimates from the Bivariate Probit Models for Estimating Factors
Affecting Participation in Tennessee Farm Fresh and Pick Tennessee Products
Parameter Estimates for the Bivariate Probit Model
Participation Equations
Independent Variablesa
Tennessee Farm Fresh
Pick Tennessee Products
-1.4138
-1.2286
Constant
(1.5797)
(1.1598)
0.0072
-0.0056
AGE
(0.0180)b
(0.0123)
-0.1354
0.5639*
BACH_GRAD
(0.3725)
(0.3178)
0.2490
-0.3815
PF_INCOME
(0.4255)
(0.3365)
0.0309**
0.0167
VEGSIZE
(0.0146)
(0.0112)
-0.0071
-0.0022
TDS
(0.0074)
(0.0065)
-0.0238
-0.1986
EASTTENN
(0.4187)
(0.3344)
0.2426
0.7825
WESTTENN
(0.5254)
(0.5024)
0.0705
0.1285
BENEFIT_SALES
(0.1943)
(0.1434)
0.0483
0.3013
METRO
(0.3467)
(0.3038)
0.1841***
0.1485**
EDUC_EVENTS
(0.0612)
(0.0589)
*
0.7145
0.2200
PUBLICATIONS
(0.4217)
(0.3529)
-0.9879**
FRESH
(0.4851)
Likelihood value
-81.3887
Likelihood ratio
57.6400***
Correlation coefficient
0.7826***
a
For variable definitions see Table 6.
b
Numbers in parenthesis are standard errors.
*, **, and *** represent statistical significance at 10%, 5%, and 1% levels, respectively.
73
Table 11. Effect of the Independent Variables on the Conditional Probability of Participation in
Tennessee Farm Fresh and Pick Tennessee Products
Marginal Effects of the Bivariate Probit Model a
PART_TFF=1/ PART_TFF=0, PART_TFF=1,
PART_PTP=1
PART_PTP=1 PART_PTP=1
0.0050
-0.0033
0.0011
AGE
a
(0.0084)
(0.0045)
(0.0036)
-0.2170
0.2107**
0.0008
BACH_GRAD
(0.1938)
(0.1011)
(0.0774)
0.2197
-0.1756
0.0303
PF_INCOME
(0.2024)
(0.1140)
(0.0857)
0.0109
-0.0002
0.0066**
VEGSIZE
(0.0070)
(0.0041)
(0.0032)
-0.0029
0.0006
-0.0014
TDS
(0.0033)
(0.0021)
(0.0015)
0.0413
-0.0620
-0.0131
EASTTENN
(0.1978)
(0.1101)
(0.0827)
-0.0581
0.2255
0.0789
WESTTENN
(0.2178)
(0.1821)
(0.1446)
0.0011
0.0303
0.0188
BENEFIT_SALES
(0.0910)
(0.0495)
(0.0393)
-0.0555
0.0927
0.0219
METRO
(0.1593)
(0.0955)
(0.0715)
0.0521*
0.0155
0.0412***
EDUC_EVENTS
(0.0271)
(0.0195)
(0.0146)
0.2877
-0.0583
0.1421*
PUBLICATIONS
(0.1756)
(0.1127)
(0.0832)
-0.4841**
0.1861**
-0.1861**
FRESH
(0.1888)
(0.0769)
(0.0769)
*, **, *** denotes significance at the 10%, 5%, and 1% levels respectively.
a
Standard errors are in parenthesis.
Independent Variables
74
50%
45%
40%
35%
Vegetable and melon farmers
Census 2007
30%
25%
20%
Fruit and tree nut farmers
Census 2007
15%
Sample data
10%
5%
0%
Under 25 25 to 34 35 to 44 45 to 54 55 to 64
65 and
over
Figure 2. Age distribution of sample data compared with the 2007 Census of Agriculture
75
Part 4: Summary
76
Summary
This study evaluates the factors affecting fruit and ve getable producer awareness and
participation in Tennessee’s state-sponsored marketing programs, Pick Tennessee Products
(PTP) and Tennessee Farm Fresh (TFF). The first essay of the thesis focuses on the factors
affecting producer awareness of these programs. Univariate t-tests were used to examine
differences among producers who were aware of these programs and those not aware. Producers
aware of TFF and PTP tended to have a higher percentage of household income coming from
farming, used more University/Extension educational tools to obtain information on how to
better market produce (e.g. educational meetings, publications), sold more products in
neighboring counties and elsewhere in the state, and had more acres in fruit and vegetable
production than producers unaware of the programs.
A bivariate probit model was used to evaluate the impact of producer, farm, and county
characteristics on awareness of TFF and PTP. It was hypothesized that the error terms between
the awareness equation for TFF and PTP were correlated. The correlation coefficient between the
two equations was positive and significant, therefore, the bivariate probit approach was deemed
appropriate. The results from this analysis showed that the use of University/Extension
publications, size of the fruit and vegetable operation, percentage of sales from direct-toconsumers outlets, and location in a metropolitan county all significantly affected awareness of
the TFF program. Attendance at University/Extension education events, age, education, and
percentage of income from farming were factors significantly affecting producer awareness of
the PTP program.
77
The second essay examines factors affecting Tennessee fruit and vegetable producer
participation in TFF and PTP. The first essay provided a starting point for selecting variables
hypothesized to affect participation. Univariate t-tests were performed on the selected producer,
farm, and county characteristics in order to examine differences between producers participating
in the state marketing programs and non-participants. TFF participants tended to have larger fruit
and vegetable operations, attended more University/Extension educational events and used more
University/Extension publications to obtain information on how to better market their produce
when compared with TFF non-participants. Producers participating in PTP tended to be more
educated, have higher percentages of household income coming from farming, attended more
University/Extension educational events, and used more University/Extension publications.
A bivariate probit model was used to evaluate the producer, farm, and county
characteristics that affect participation in TFF and PTP. It was hypothesized that correlation
existed between the error terms in the participation equations for TFF and PTP. The correlation
coefficient for the two equations was positive and statistically significant which supported the
hypothesis that unobserved factors affecting participation in TFF and PTP are correlated in some
way, and a bivariate probit model was appropriate for this analysis. University/Extension
educational events and whether the producer sold fresh fruits and vegetables had significant
impact on participation in TFF, given producer participation in PTP. Factors that significantly
affected the probability of participating in PTP, when not participating in TFF, were whether the
producer had a bachelor’s or graduate degree and whether the producer sold fresh fruits and
vegetables. Size of fruit and vegetable operation, University/Extension educational events, the
use of University/Extension publications, and whether the producer sold fresh fruits and
78
vegetables were significant factors affecting the probability of jointly participating in TFF and
PTP. It is interesting to note that producers who sell fresh fruits and vegetables and participate in
PTP are less likely to also participate in TFF, regardless of the fact that TFF specializes in fresh
produce.
The information gained in this study makes a significant contribution to the body of
literature concerning producer awareness and participation in state-sponsored marketing
programs given the limited number of studies currently available (Govindasamy et al. 1998a).
This information can benefit policy makers in the state of Tennessee, such as the Tennessee
Department of Agriculture, as well as policy makers in other states. By identifying the profile of
farmers who are aware of and are participating in these types of programs policy makers can
modify programs in a way that they can increase awareness and participation, better address
clientele needs, and adjust limited funding.
The findings contained in this research may also benefit University/Extension programs.
The attendance of University/Extension educational events and the use of University/Extension
publications to obtain information on how to better market produce were significant factors
affecting both awareness and participation in TFF and PTP. These results highlight the
importance of continued partnership between policy makers and University/Extension personnel
in order to increase producer awareness and participation.
Further research is needed regarding consumer perceptions and preferences for produce
labeled with TFF and PTP marketing logos, as well as consumer willingness to pay for products
marketed through these programs to evaluate effectiveness of the programs from the consumer
79
stand point. With a more complete picture of effectiveness of state-sponsored marketing
programs in Tennessee policymakers may be able to make informed decisions regarding
improvements or discontinuation of state-sponsored marketing programs.
80
Vita
James Andrew Davis was born in Cookeville, Tennessee on September 7, 1984, to
Wayne and Tana Davis. He graduated from White County High School in Sparta, Tennessee in
2002. He attended Tennessee Technological University and received a B.S. degree in Agriculture
with an area in Agribusiness Management in December, 2009. He later attended the University
of Tennessee, Knoxville, where he earned a M.S. degree in Agricultural Economics in May,
2012.
81